19 May 2008

Revisiting the Season Prediction

Several weeks back I predicted the number of runs the Orioles would give up and the number of runs the Orioles would score. The basis of this prediction depended on a few assumptions:

1) ZiPS/Morong Formula (my arrangement) would properly predict offensive and pitching performance.
2) Offensive replacements would cause a 10% reduction in run scoring while unearned runs would be ignored for pitchers.
3) Top 5 starters would start every game and provide an average of 6 innings pitched.
4) Relief pitchers would be league average.

1. ZiPS/Morong predicting performance.
ZiPS actually overpredicted the runs scored (with the run reduction application). ZiPS predicted that 193 runs would be scored. In actuality, 179 were scored. Even more of an issue was prediction of pitching performance. 222 runs were predicted, while 184 were actually scored. A major contribution to this error was the unexpected development of Daniel Cabrera and a bullpen that was much better than expected. It should be mentioned that my placeholder of a league average bullpen was actually somewhat optimistic. This formula under predicted the Orioles success.

2. Offensive reduction and static pitching.
My educated guess of a 10% reduction was pretty apt. Plugging in the actual OBP and SLG of each player resulted in a coefficient of 0.927 to reach the actual runs scored. The pitching prediction appeared a bit too kind. After plugging in the actual SP and RP era, the system predicted 177 runs, where there were actually 174 runs scored. The application of a coefficient (1.057) would have been appropriate to account for unearned runs.

3. Top 5 Starters would remain so and would average 6 IP.
It was to be expected that a starter or two would be injured. It was known this was a weak assumption. Loewen's injury made it so. The 6 IP prediction is actually almost right on the button.

4. RP would be league average.
Orioles RP are actually pitching 13% better than the league average bullpen.

New Adjustments


1. ZiPS is being replaced by PrOPS and xFIP.
As the season continues, in-season statistical methods may actually predict future performance better than season beginning predictions. The reason for this is that certain growth or degradation may not be apparent prior to the season. PrOPS takes in peripheral batting data to predict OBP and SLG. xFIP takes peripheral pitching performance data and predicts future ERA. Current relief pitching ERA will be multiplied by the coefficient factor mentioned in the next paragraph.

2. Performance Coefficients

The batting performance reduction coefficient will be changed from 0.9 to 0.927. The pitching performance coefficient will be changed from 1 to 1.057.

3. Record Calculation
The current record is considered a given, so the new predicted winning percentage will be applied to games yet to be played. The number of wins determined by the formula will then be added to the current total.

Team Used for Calculations
PrOPS/Morong
2 Roberts.........374obp/424slg
3 Mora............348/454
R Markakis........414/521
D Huff............324/441
L Scott...........328/395
1 Millar..........354/444
C Hernandez.......326/332
C Jones...........293/354
S Placeholder.....300/330

xFIP
S Olson...........3.75
S Guthrie.........4.19
S Cabrera.........4.19
S Trachsel........5.92
S Burress.........4.64
R Bullpen.........3.42

Results

In the games left, this method predicts the Orioles will score 542 runs and give up 537 runs. The season ending run totals would be 722 runs scored and 721 runs given up. In the remaining games, the winning percentage would be .505, which would end with us having a .518 winning percentage at the end of the year (Pythagorean Win Expectancy). This means that the current rendition of the model predicts we wind up with a 84-78 record, which is 2 wins above the PWE.

Discussion


The current model is placing a great amount of worth on the ability of PrOPS and xFIP to accurately predict future performance. In addition, the new coefficients are assumed to remain constant. Finally, expecting the current rotation to remain as the final rotation is, again, quite a weak assumption. Anyway, things look a lot brighter for the O's than it did a few months back.

15 May 2008

What to do about Orioles' Catchers

A growing conversation in the past few weeks has been about the Orioles' catching position. It was thought this year was going to be a lost year and that catcher wouldn't be a major consideration (or shortstop for that matter) even in light of a horrific offensive showing. The two Orioles on the major league squad manning the position are Ramon Hernandez and Guillermo Quiroz. The guy most talked about replacing these folks is last year's 1st round pick, Matt Weiters.

Ramon Hernandez
Ramon Hernandez is an injured and undependable catcher. His inability to block pitches and react to poorly located balls often gives spectators the idea that he is nonchalant in his work. That really is far from the truth, but it seems to make people feel better if they can turn physical limitations into mental ones. People can be blamed more for the latter. What I think we are observing is the physical decline of Hernandez. We signed him at age 30, which is old for a catcher. Particularly, for a catcher's defense. Age usually results in more injuries and four out of his past five years have been involved at least one major injury that sidelined him. Of course, we signed him less for being a better defensive option that the old and fading Javy Lopes (and Ramon was better defensively than Javy), but also because Ramon was an above average offensive catcher. His first season he was well above average (an OPS of 90 is average for a catcher), he second year was average, and this current one has been poor. I fear that this might be Ramon's last multiyear contract. I think it is likely that he does not rebound offensively or defensively . . . and a catcher with poor motor response who cannot hit is only worth so much. He does seem to be pulling the ball more than he has in years past. I wonder if he is cheating on fastballs.

Guillermo Quiroz
Quiroz was a highly touted prospect in the Blue Jays system. Well, not at first. He was signed as a 17yo international free agent. His defense was excellent and much of his minor league career he average about 0.12 PB/9. That is outstanding. His 4th season he was promoted to HiA and began to display some power as he jumped from 21 to 41 extra base hits. The following season at AA, he broke out with 890 OPS (he was 21yo). He tried to play through a collapsed lung, but had to be shut down as he could no longer run. As he progressed, he began to lose the plate discipline he showed in AA and his defense started to erode. Part of that was blamed on a broken hand. Baltimore is now his fourth organization and his future does not look much brighter than being a backup catcher. He still flashes power and could be something if his contact rate improves, but he has shown to be a target for opposing base stealers to take advantage of. He will probably turn out to be slightly better than Geronimo Gil (how did that guy ever get to start?). He is fine as a backup, but nothing more.



Matt Weiters
Weiters was the 5th overall pick in the 2007 MLB draft. He fell that far because of fears of Boras' demands. Weiters proved difficult to sign, but right before the deadline his father called Peter Angelos and was able to get the ball rolling. He didn't play short season last year, but was impressive in Hawaii*. This year he was slotted in to the lineup in Frederick where he has proceeded to deliver a 333/421/582 line, which is a 136 OPS+ for the Carolina League. His equivalent line in the majors would be 258/319/426, which is a 102 OPS+. Right now, Wieters is an above average offensive catcher. Right now, his defense is probably on par with Ramon and maybe on par with Quiroz. Right now, Weiters is the best catcher in the organization. That is a problem.

The Problem
There are a few reasons for keeping Weiters in the minors. First, he may have holes in his offensive game that will be easier to work on in the minors than in the majors where the quality and placement of pitchers is significantly better (remember Weiters is in HiA ball). Second, Weiters also will have to learn at the MLB level how to call a game, which probably takes some time. Third, this will be Weiters longest and most grueling season ever. To add the physical strain of being a catcher to the mental strain of adjusting to the MLB game . . . it is asking a lot.



Suggestion
Weiters probably develops just as well in HiA. In another month, promote him to Bowie and let him ride the year out there. His bat is looking good enough that I would start eyeing another position for him to play when he gets to be an amazing hitter and a poor catcher (around age 27 or 28). Ramon needs to go. If his value can rise high enough, send him off this summer. If not, try to package him this winter. The problem with this winter is that there are a few guys like Hernandez out there who will take 1 year deals for significant less money (Barajas, Estrada, Barrett, Irod, LoDuca). Last offseason would have been better. We may not be regretting the loss of Lastings Milledge, but I imagine we could have gotten something in exchange for salary relief and similar production from some other retread catcher.

Wieters' Future
I have thought of him as a catcher until about a month ago. His hitting is just unbelievable. He could be a truly special player at several positions. I would start him out in the majors as a catcher, but would give him some looks at other positions when he felt more comfortable. That being said, if Tim Beckham or Pedro Alvarez is not available . . . Posey would be a nice acquisition with the 4th pick in the draft. Posey would be MLB ready about the time it would be a good idea to shift Wieters to another position.

14 May 2008

Daniel Cabrera's Batted Balls


Daniel Cabrera pitches tonight against the typically patient and talented Red Sox offense. He has been able to string together six seemingly good games in a row. Several of his game scores were not impressive during this period as three were in the 50s. This helps give the impression that he has changed and is a much better pitcher. In fact, quotes are beginning to be dug up by unnamed scouts declaring the Cabrera has turned the corner and is now a legitimately fulfilled talent. Now, he might be. He might actually be a very good pitcher now. He certainly is throwing differently as in he has really imploded in any start this year other than his first. He has certainly comes close several times during this stretch though. I mentioned in a previous post how his peripheral numbers do not make any sense and that he is more truly a 3 or 4 and not an ace as his current ERA suggests. Today we are going to look more into his batted ball data.

Battered Balls
I have taken his batted ball data from baseball reference and normalized it to the 2007 AL batted ball data. As you may realize, changes in batted ball data can be explained by: 1) changed pitch run (this is quite rare for an established pitcher to drastically change how his pitches run to the point that his batted ball data would also change dramatically), 2) improved defense (this can greatly improve or hinder BABIP, but typically teams do not make wholesale changes in defense from year to year), and 3) pure luck (batted balls are typically not uniform in their dispersion over the course of partial seasons and sometimes full seasons). Using that base knowledge, we are able to discern potential discrepancies in Cabrera's batted ball data.

Basics
Infield and Outfield - BABIP of balls hit to these respective areas
BABIP - Overall batting average of balls in play
First, Middle, and Third - BABIP of balls hit to these areas of the field, including the outfield
GB, FB, LD - BABIP of balls classified as groundballs, fly balls, and line drives.

Results
When you look at a graph like this, you look for extreme differences. A difference of about 20% is a good rule of thumb for identifying extremes. Using that as a benchmark we have these weird BABIPs: overall BABIP, infield, third, middle, groundballs, and line drives. It is good to categorize the data because it can give us better ways to identify what is going on here. BABIP alone shows that something weird is going on, but the rest of the data suggests why.

Fielding cannot really be relied on as having changed. Much of the difference looks like it applies to groundballs going up the middle and to the right as well as line drives in the same area. The only difference in the infield is Luis Hernandez and he got benched because of poor defense. There are seems to be little reason to think that Brian Roberts or Melvin Mora have been significantly better at what they do this year. In the outfield, we have Adam Jones and Luke Scott. I think we can say Luke Scott improved our defense. I also think Adam Jones is better than Corey Patterson. I also think the differences between this year and last is just not enough to explain the differences. What is telling is that difference in line drive percentage. You can explain away grounders by suggesting Cabrera is inducing weak groundballs, which is plausible (though I doubt with his repertoire of four seamers--it should be noted that althought I see little difference between pitch quality this year and last beyond placement . . . I could certainly be wrong about the quality of his stuff compared to last year). You really can't explain his dramatic difference in line drives. Line drive success has little to do with good defense and more to do with simple luck. Not only is his line drive percentage unsustainably depressed . . . the line drives that are actually hit off him are going straight to his defenders at an unsustainable rate. Add this to my doubts about the groundballs being more effectively covered and you have a guy who is ripe for a downturn in success.

Conclusion
I will be bold and predict that Cabrera has trouble tonight. Some people will point to his 116 pitches in his last outing. Others will just say he is no good. Or maybe he will balance a dozen plates and toss 8 innings and give up a run or two. My guess though is that he will eventually drop down to about an era in the high 4s by the end of the season. He has been lucky avoiding baserunners and getting out of jams. His LOB% is about 10% above where it should be. His K/BB is 1 when people are on base, which is a third worse than league average. There just isn't much to like with him going up against a team with the 3rd most walks, best obp, and the best slg in the AL. Anyway, cross your fingers

13 May 2008

2008 Playoff Odds Update

I have added a new feature to this weekly column: the Pythagorean Expected Wins. This method is often scrutinized, so I felt it would be interesting to throw it up there and see how it responds throughout the season. This method will not predict playoff chances. I have also narrowed the viewable field. I will continue to do this as the season goes on in order for us to more easily discern the differences between the methods presented.

PECOTA has downgraded the Orioles' chances from 1:176 to 1:205. ELO has actually upgraded us from 1:16 to 1:15.

12 May 2008

Napkin Scratches: Handle with too much care?


One topic that comes up time and time again is how strong and durable pitchers used to be. Gone are the days where a pitcher would finish half the games he started. Gone are the days that no one needed a LOOGY. These are things we often lament. Right in hand with that lamenting is a palpable anger toward today's pampered pitchers. These guys are given strict pitch counts in the minors right into the majors. Gone too are the days when it was common to hear about some obscure prospect in PCL who threw 255 pitches in a single game that lasted 18 innings. It just isn't done.

So . . . why?

The first culprit is money. In 1976, the Reserve Clause was shot down. This enabled free agency and it changed the way a lot of different aspects in baseball were handled. In terms of pitching, it affected a few things. First off, it increased the value of pitching. Not only did the current game or season matter, but also seasons down the road. You had to invest a significant amount of money into a pitcher and you would not want to take a loss on it. This may have been the main pushing force of the five man rotation. Although, the five man rotation was not unheard of before the ruling, it quickly became the norm after it. Of course, as you may remember, the Orioles switched to the five man rotation in 1983 when Joe Altobelli took over.

Now, what the five man rotation does is give the pitcher more time off between starts for rest. The idea behind this is that pitching is a violent action. It is also not a very normal way to use your body. Quite unnatural. It follows reason that chronic repetition would result in physical damage. More time off between starts should equate to allowing the body to recuperate. Others argue against this though because it results in an uneven workout schedule. In the 4 man rotation, you would take a day off, throw a side session, then take a day off between starts. There is really little reason to think an uneven schedule hurts performance for any athlete that is not obsessive compulsive about such things. In my opinion, the value of a 5 man rotation is only a plus if the amount of strain forced on a pitcher requires an extra day of rest. It should also be expressed that not all arms are alike.

After the five man rotation was cemented in, pitch counts started popping up on the horizon. As far as I can tell, concern about pitch counts emerged in the late 80s and early 90s as drafted pitching prospects began making serious mint. The philosophy seeped into the majors in order to help preserve oft-injured pitchers (i.e. Saberhagen) or with fresh faces (i.e. Josh Beckett). The number typically chosen is 100. The PAP system goes with that. Top tier pitchers usually throw about 100 pitches per game. Even guys who claim not to use counts (i.e. Oziie Guillien) still seem to have pitchers that hover around the 100 mark.

That is the weird thing. Coaches who believe in pitch counts and those who don't . . . they do not vary much in terms of when a pitcher should be taken out. Again, this is now a bright line criteria. There is variation from arm to arm, some seem to be able to handle more and some less. This is pretty much regardless of coach. It seems that either no one uses pitch counts or that pitch counts just don't vary from normal performance measures.

This got me to wonder some things:
1. Training and development is better now than it was in the past
2. Players are more athletic and stronger now than they were
3. Pitching, which is more reliant on tendons and ligaments, has less potential to benefit from improvements in other areas of training and sport medicine. For instance, you can strengthen a pitcher's legs, but you cannot strengthen his tendons and ligaments to withstand a higher degree of torque.


So:
As technology and knowledge increases, so does the player's ability to perform, which is most likely to benefit a hitter more so than a pitcher due to the physical limitations placed on each activity.

That reminded me of this study. Read it. Seriously, read it. Good stuff. It is not perfect, but it provides a great approximation of how developed talent has increased over the years. Some interesting things pop out in that study. The ten best hitters ever in order from best to worst:
Bonds, Ted Williams, Aaron, Musial, Mays, Frank Robinson, Yaz, Ricky Henderson, Cobb, and Mantle.
Honestly, it makes sense. People often overlook how good guys like Ricky and Musial were. It is also interesting to think that Ted Williams would still kill in today's game.

Anyway, if players have gotten better . . . wouldn't pitching have gotten tougher if you assume that hitters will benefit more due to not being so reliant on tendons and ligaments? To this extent, we can use league quality as a coefficient to determine how today's pitching load per start compares to other eras. For this napkin scratching effort I am considering all starting pitching data in the AL from 1969 to 2007. I am predicting pitch counts based on the method I previously used. I am also normalizing the pitch counts based on the competition level of the league, taken from this graph.

RESULT
The black line is raw SP Pitches per Game and the orange line is league quality adjusted SP Pitches per Game.


What is interesting to note here is that as pitch counts are reduced, the league quality coefficient somewhat accounts for it. This suggests that league quality is causing a decrease in pitching counts. Others have countered that it might just be a relationship with runs per game. So, I ran another one with runs per game and as you can see there is much more variability in the runs per game adjusted line (black) and the league quality adjusted.



Conclusion
League quality appears to be affecting the number of pitches a SP throws. This suggests that pitch counts are either not followed or are more complex than a single line. It appears as if pitchers are being used quite similarly to how they were back in the 70s even though IP/G has dropped from 6.2 in 1973 to almost 5.2 in 2007.

09 May 2008

Cabrera Strikes Again


Daniel Cabrera has been able to pitch well on a rather consistent basis using a fastball. This sort of thing is possible in Little League, but it should not be in the Majors. Particularly, he throws it in the low 90s. He benefits from a BABIP of .226 while his career BABIP is .300. His FIP is 4.75, while he ERA is 3.54. He is posting his worst K/BB rate for his career (1.43) and has upped his career left on base percentage by ten points to 78.9%. All of this marks for a downturn. The only thing that explains it to me is that his Line Drive percentage is at 11.8%, which is about 6% less than he normally posts. As you all know, a rough hand estimate for batting average is taking your line drive percentage and adding 12% to it. So, with an 11.8%, you'd explain a batting average of roughly .238 and that is essentially what we have. His groundball percentage is also 6% higher, so it seems he has essentially traded some line drives into groundballs. The weird thing here is that his fastball looks like a four seemer as opposed to a sinker. How is this possible when you only throw fastballs (85% of his pitches are fastballs) and they average at 92.9 mph? I don't know, but that graph below is pretty (note: he has little command of his slider).


Cabrera's Fastball and Slider Pitch Location for May 8, 2008 against the Royals.

08 May 2008

Adam Jones Spotlight


Adam Jones has had a rough year so far. He has hit 231/272/359. This was expected. He is a contact-based hitter (meaning he does not and most likely will not walk much), so until he understands how to hit off speed pitches . . . he will be taken advantage of. There are some good things to like. Defensively, Jones has been quite good in the field. He has saved about 1.5 runs (+2 plays and 1 assist). He also has made 19 plays out of his zone, which is tied for the AL lead with BJ Upton and Torii Hunter. I imagine over the course of the season he will wind up being about +4 or +5 runs in the field, which is good for about a half win above median.

Offensively, well, we'll take the good with the bad here. Let's go into some Pitch f/x data.

Pitch Types

Fastball

Many folks want to place Jones in the two spot in front of Markakis. This would enable him to see more fastballs than when he bats in front of Luis Hernandez or some other below average hitters. At least, this is the hypothesis people use. Jones has seen 222 fastballs, which comprise 60% of the pitches he sees (this compares to 67% Mora sees in the 2-spot and league average is 58%). It doesn't seem like he is being thrown an abnormal number of off speed pitches. I also wonder to what extent Mora's 67% is a product of his declining bat speed because last year he was in a similar lineup position and saw the league average of fastballs. Anyway, lefties and righties throw him about the same number of fastballs. Two-thirds of his at bats end on a fastball. As you can see on his fastball graph . . . he goes after the inside pitches, but lays off the outside ones. Two guesses on this one: 1) he is trying to push the count and try to walk or 2) he is not an outside fastball hitter and is waiting for a pitch in his zone at which to swing. You may also note that when he does swing at pitches on the outside he makes contact. Every time. He has no swinging strikes out there, but when he chooses to swing he has four singles and a lot of fouls and in play outs. I guess the bottom line to take from this is that it appears he has a plan, which is good . . .usually.

Curveball

Jones has seen 25 curves or about 6% of the pitches he has seen. Lefties and righties throw it equally and the percent is a third less than the league average. Only 3% if his at bats end on a curveball. As the graph shows . . . he swung for 12 of them, making contact with seven. Six of those balls were not in the strike zone. He also chose not to swing at five that were strikes. In terms of pure pitch recognition . . . he "saw" 14 of 25 as their correct calls (56%) assuming he did not choose not to swing at a pitch he knew to be a strike. Torii Hunter has a 50% recognition on curves this year.



Slider

So far, righties like to toss sliders at Jones. He has seen 65 (18%), which is 20% more than league average (this may be the result of us facing more right handed pitchers, but I am not sure about that being true). Almost 22% of his at bats end on this pitch. He displayed a 78% accuracy rate in reading these pitches with 6 called strikes. Torii identifies 62% correctly with 5 called strikes. Again this "accuracy" rate assumes the players are not choosing to ignore strikes. Jones seems more adept here.




Change Up

Lefties love to toss the change to right handed Jones. Again, this is probably due to it being harder to hit a pitch breaking away from the batter than a pitch breaking into the batter as evidenced by personal experience, watching, and from several physics papers. A southpaw is three times as likely to throw a changeup to Jones than a righties. His overall rate is 10%, which is a sixth less than league average (probably due to my unresearched belief that we have faced more righties than the average hitter). Only 4% of these pitches end his at bat. He shows 68% recognition with five called strikes. He seems to be able to recognize these pitches as strikes, but misses them. This seems like it should be a pitch people could rely on to get him out as it appears he thinks a lot of these are fastballs. Torii recognizes 21% of these and hits them well. With a similar number of pitches, Jones is three times more likely to swing and miss a change up. This seems like an area he will need to improve upon if he wishes to hit better against lefthanded batters.

Pitch Counts
It should be noted that getting behind in a counter from the hitter's perspective is the opposite of that with the pitcher. So, if you want to compare this chart to the pitching version, flip them. Anyway, what we see here is that Jones, as most batters, gets behind in the count often. He is not as successfully aggressive as Jones is and is not as patient as Markakis is. Pitch counts for batters is not as effective as a measure as it is for pitchers because the data will skew much more due to individual batting approaches. What it does show is that, along with the Pitch f/x data, Jones needs to do more with his batted balls.

Conclusion
Jones is struggling like you would expect a young contact-based hitter to struggle. He is learning to identify much better off speed pitches than he faced in AAA and that takes time. Torii Hunter's career is a good comparison. He struggled early and as he learned to identify pitches . . . he has gotten a lot better. Adam Jones' approach may be similar to his, but I think Jones has more power potential and not as much defensive potential. That said, enjoy these moments as we watch a very good talent adjust his game.

07 May 2008

Garrett Olson's 2nd Game and Pitch f/x


So, yeah, I wrote that Garrett Olson was a prime candidate to get shelled in his next game when his unsustainable LD% becomes sustainable. I mentioned how his pitch counts were also going to hurt him. So what does he do? He 4 hits (w/ no extra base hits) the wild card leader and marks 7 K's and a single walk (to his last batter). Now, the caveats to this performance are: 1) the A's have no power and rank 13th in slugging in the AL, 2) Oakland Coliseum is a pitcher's park, 3) Frank Thomas would have had a homerun in almost any other park, and 4) Garrett is somewhat lucky. Now, that said . . . he did a great job with his pitching counts and seemed to be quite aggressive toward the batter. This is what he needed to do and he did it. I don't think he is an ace type of pitcher, but I think he can deliver a 100 ERA+ or a little below. So I'll go through some of the data from the Pitch f/x system:

Pitch Type


More curveballs peeked into the Pitch f/x data. I still don't know what I happening. I thought Olson threw a fastball, change, and curveball. It may be that the equipment here are recording lower speeds than those at Camden Yards and the logarithm that predicts pitch type is seeing the lower speed and calling more of them curveballs.

Fastballs appear to be thrown at the same frequency between the two starts with 57%. Changeups have increased from 10% to 18.75%, which is most likely due to the lack of left-handed batters. Garrett does not throw his change up to lefties. The curve/slider has decreased from 38% to about 24%. This is again most likely due to the heavily right-handed lineup Oakland uses.

All in all, 55 fastballs were thrown (69% strike or hit in play), 18 changeups (60%), and 23 slider/curves (70%). In terms of purely missing the ball, the harder curveball appears to be his major out pitch. He seems to have less control of the pitch as it decreases in speed (likewise, increasing in drop). The graph to the right displays strikes, ball, and hits in play.


Location


Lefties
Here are his location and pitch types (green=fastball, orange=slider, brown?=curve) for lefties. The first thing you probably noticed is that Oakland has far fewer lefthanders in their lineup (only 1). The next thing you probably notice is that Olson hit his spots against Oakland. He really is hugging the outside part of the plate. When you see this sort of difference between starts . . . you get led to release point and pitch breaks. The pitch breaks look the same, but the release point reveals a major difference between the two starts. On the April 29th game, his fastball release point varied horizontally by almost a foot. This looked a lot like Cabrera's typical release point. Against the Athletics, Olson actually narrowed his release point to a range of about 6 inches. I think him being able to position the ball is linked to that metric. That metric suggests that his delivery was more repeatable than it was in his 2008 debut. Oh, right, this is about left handed hitters. Not much can be said as he faced so few in his second start. He still isn't throwing his changeup to them. Three of his walks came against lefties in the debut and his only walk came against a lefties in this most current game. An issue? I don't think we have enough data on that yet. Last year, he walked lefties and righties at the same rate.

Righties
Olson's first game consisted of pitching insides as often as possible. It looks like he had more spray to his location. In Oakland, he did not focus inside much, but was incredibly sharp at keeping the ball in the lower 2/3rds of the strike zone. This may be the result of better command or a more uniform approach to each hitter. I'm not sure there is much else to discuss on how he pitched to righties. Feel free to comment if you see something.

Speed
Speeds were about 3-4 mph less than his debut. This may just be a difference in equipment between stadiums or he could have toned it down a bit. I'd bet on the former. Also, more of an appearance of the curves and sliders.



Pitch Counts
One of the problems I saw with Garrett last start was his inability to get ahead in his pitch counts. He was actually worse against the Rays than he was during his forgettable 2007 run. It was one of the major reasons why I thought he would get shelled. If you combine a lot of guys walked on base and an average line drive percentage (more on that in a moment) . . . you are going to get hit. Hard. Anyway, he turned it around rather dramatically. His behind counts went from 27% to 10%. His even counts went from 31% to 24%. His ahead counts went from 42% to 67%. Typically, when you are ahead in the count 67% of the time . . . you are going to do well. I wonder if the A's approach of only swinging at their pitches allows pitchers to get into favorable counts. I'm not sure. Anyway, he did a great job reversing these numbers from last time. League averages are 12% for behind counts, 29% for even counts, and 59% for ahead counts. His numbers are potentially sustainable. It also sheds light to how poor his numbers were before.

Hit Quality
The second part of my hypothesis about Olson getting shelled was that a 7% LD rate is unsustainable. I swear this line drive percentage is not sustainable. I'm not aware of another pitcher who is capable of putting up such a line. This is just crazy. So, yeah, 7% again. His groundball rate dropped to a more realistic, in my opinion, 42%. His left on base percent remains in the high 70s. So . . . either his stuff is nasty and no one can see that . . . or he benefited from a weak hitting Oakland lineup who somehow couldn't figure out how to earn a walk against him (A's are 3rd in runs scored).

Conclusion

That line drive rate will go up. It won't hurt him much if he can continue to not walk anyone. I'm still calling for him being a 3 (95 ERA+) or 4 (85 ERA+) this year. I guess we will see.

06 May 2008

Have we passed the high water mark for 2008?

This week is less rosier than last week. PECOTA downgrades our playoff chances from 100:1 to 176:1. ELO, though more optimistic, agrees that this past week did not bode well for the plucky Orioles. They dropped them from 8:1 to 16:1. After going 2-4 since our last playoff odds update . . . the O's are sitting at 16-16. Next week I will update the ZiPS/Morong derived prediction. Oh . . . here is the chart:



NOTE: I am going to try to quickly write up a Garrett Olson piece tonight. Just skimming the data on Pitch f/x . . . looks like a much better pitcher. I also think his counts were in better shape. Anyway, I love it when I project some negative criticism toward an Oriole and get proven wrong.

05 May 2008

Shortstops and 2009


As mentioned last week, Luis Hernandez has been awful. He was brought in to play defense and anything resembling offense would be appreciated. Well, according to RZR . . . he is the worst defensive SS in the AL. His offense 241/303/259 would inspire a Woody Guthrie tune about the collapse of production. It is not good and he cannot remain our shortstop. There is really no scrap of evidence that has been presented that shows he deserves to be above AAA, in my opinion. At AAA, he would be a league average hitter with decent defense. At the MLB level, he is one of the worst offensive players ever and a poor defensive shortstop this year. This entry will try to determine if there are any better in house options or players potentially available this offseason. If anyone wishes me to add anyone, let me know.

Method
Using Cyril Morong's run expectancy formulas, I compared each player as potential 7, 8, or 9 lineup position hitters. The algibraic mean of their expected run performance was used and related to Luis Hernandez'. Defense was based on RZR. If defensive statistics existed for SS, an approximation of expected RZR was used. If not, below average defense was determined to be 7% worse than the average. Performance data for players with less than a season of information was taken from their accumulated experience. Players with more than a year of experience used ZiPS projections for offensive performance and an average of their last three years of defensive performance at SS. Due to the ages of most of the players, the predicted defensive performance is most likely a liberal estimate as SS play deteriorates rapidly with age (more so than with any other position, including catching). As we had done before, these are converted into saved and given runs. It was assumed that SS face 400 chances over 162 games. These numbers are converted to wins/162 games based on dividing runs produced and runs saved by 10.

Results

Luis Hernandez and the Average SS
Luis Hernandez, based on his career, is a bottom of the latrine hitter and a good fielder. The average SS in 2007 had a batting line with a .330 OBP and .407 SLG. The average SS also had a .807 RZR. Luis' career .301 OBP and .315 SLG results in 21 less expected runs, which is divided by 10 to give -2.1 wins (compared to the average offensive production for a SS. His defense has been measured to be 0.842 RZR (we'll assume his career RZR is more accurate to his true production at SS than his 2008 .765 RZR). His defense is worth 0.8 runs above average or +0.8 runs. This reduces his cost to -1.2 wins (after rounding). This means that having Luis as our shortstop will cost us 1.2 wins in comparison to having an average SS in place. If he truly had the .930 RZR he showed in limited play last year, he would actually have been worth 0.9 wins . . . that is a difference of 2.1 wins. Using the same reason, this year's .765 RZR would mean his worth would be -3.1 wins. It should be noted that for the purpose of this study, we are using the "average" SS. Typically, replacement level performance is used. A player worth 2 wins below average is about what a replacement level player would be worth. Replacement level as I define it is your average AAA SS.

In-House Options
The in-house options are Eider Torres (.282 OBP/.314 SLG, 0.800 RZR; -2.1 bw, -0.2 fw), Alex Cintron (.291/.350, .765; -1.9, -1.0), Brandon Fahey (.316/.330, .800; -1.4, -0.2), Freddie Bynum (.307/.379, .750; -1.0, -1.4), and the improbable Scott Moore (.311/.409, .750; -0.5, -1.4). Of these five in-house options, none are better than what we project for Luis Hernandez. He seems to be our best SS option. We have no potential SS who are better than -1.2 wins.

2009 SS Options
Potential Free Agent Possibilities are as follows:
David Eckstein (336/344, .840; -0.7, 0.8)
Adam Everett (285/330, .860; -2.3, 1.3)
Orlando Cabrera (332/372, .800; -0.4, -0.2)
Rafael Furcal (347/387, .830; 0.2, 0.6)
Christian Guzman (302/342, .800; -1.7, -0.2)
Felipe Lopez (345/383, .780; 0.1, -0.6)
Cesar Izturis (296/318, .850; -2.2, 1.0)
Juan Uribe (304/427, .800; -0.5, -0.2)
Of these options, only Cristian Guzman comes off worse than Luis Hernandez and he does it by half a win (5 runs). Now, you may be shocked here, but we are using ZiPS as the measure of true performance. Guzman's recent play may be legit . . . if so, he will be worth 0.1 wins instead of -1.9 wins. The only free agent shortstops that are projected to be better than the average MLB SS are David Eckstein and Rafael Furcal. Eckstein is better by 0.1 wins and Furcal by 0.8 wins.

Discussion

How much does a win cost?

A rough figure of 4.5 MM has been mentioned as how much a win costs on the open market. This typically ignores the reality that improving by a win is more important for a 90 win team than it is a 100 or 80 win team. What I mean is that wins are worth more to teams near the Wild Card or pennant than those who are further from that dividing line. For instance, a 90 win team signing a player who improves them by 2 wins for 12MM makes more sense than a 70 win team signing the same player for the same price. Why? Because 2 more wins will not get a 70 win team into the playoffs. Some fans demand a team win as many games as possible, but it can be easily seen that sometimes . . . that money can be better spent for production at a later date (i.e. prospects, saving up for a difference making free agent). So where does this leave the Orioles? They are a 70 win team this year for all intents and purposes. They could go as high as 80 and as low as 60. Next year, they will probably be a 75 win team. To be a playoff team we have to improve this roster by 15 wins. Furcal (+2), Teixeira (+5 guess over Millar), Sabathia (+7 guess over Trachsel), and our own player improvement (i.e., Markakis, Jones). The is about 50-65MM in annual salaries . . . maybe 400 MM in total contract money.

The difference Furcal makes in improving the Orioles offense, really doesn't mean much unless we go out and throw money around this offseason. So remember that the next time you curse Hernandez. Unless we acquire Teixeira and Sabathia . . . it probably doesn't matter.

02 May 2008

Garrett Olson's 2008 Debut and Pitch f/x

Garrett Olson pitched last year in the majors and suffered a forgettable 32.1 IP. He logged in 28 Ks, which is respectable. He also wound up with 28 walks, which is not. He was also whacked with 42 hits. It was a rather forgettable first taste of the majors and he was chased with a lot of criticism about nibbling down and away.

In spring training this year, he was touted as the most likely fifth pitcher. Instead, the team signed Steve Trachsel and Olson was sent to the minors. The Orioles pitching coach informed him that he needed to learn how to go after hitters and not pitch "scared." At Norfolk he pulled in this line:

24.1 IP 1.85 ERA 18 K 9 BB 1.36 WHIP

Without knowing much about his starts . . . that WHIP suggests to me that he is quite hittable to AAA hitters. They seem not to be able to turn those hits into runs, which means there is a lot of luck or that Olson knows how to pitch and change speeds.

This past Wednesday, Olson was called up to face the Tampa Bay Rays. He came out with the win and had this line:

6.2 IP 2 R 4 H 6 K 5 BB

That walk rate is troublesome. Doubly so when that was one of the main issues with his poor performance last year. The hit rate is nice, but one game hit rates are notorious for being misleading. It was not a dominating performance, but at his age . . . he has room to grow. It looks, for the moment, to be good enough to give him five or so starts and then reevaluate. So let's get acquainted with Olson.

Pitch Selection:
Olson appears to use four pitches based on the f/x data. He throws a fastball, curve, slider, and change up.

Fastball
His fastball tails about 9.5 inches and remains 7.9 inches up in the air (due to backspin) from a motionless pitch. It ranges from 88-93 mph. He throws this pitch 57% of the time and equally to each batter. With the small sample size of the number of pitches he threw, fastballs can be expected the majority of the time on 0-0, 1-0, 1-1, 2-0, 2-1,2-2, 3-0, and 3-1 counts. He never threw one on two 1-2 pitches and rarely threw it on 0-2 and 3-2 counts.

Curveball/Slider
I thought Olson threw a mid-70s curveball, but it looks like the f/x system is a bit confused on this. It says he threw three curveballs and twenty seven sliders. He either threw a hard curve or a soft slider as the two groups look more like subsections than different pitches. I'm going to assume that he still uses his curve, but throws it slightly harder than I remember. Anyway, speed ranged from 79-83. The release point drops slightly in comparison with the fastball. It comes in 11.6 inches below the fastball or 2.1 inches below the trajectory of a spinless ball. The pitch breaks across 2.5 inches horizontally. This sure sounds like a weak slider. This pitch is thrown 38% of the time. 80% of these pitches were thrown to lefties and it is often used as his secondary strikeout pitch with the fastball being the primary one.

Changeup
His changeup has a very similar release point to his fastball, which it should as it has to look like a fastball. Speed ranged from 79-83. It tailed slightly more than the fastball at 11.39 inches and came in about the same height. He threw this pitch 10% of the time and only to right-handed batters. He preferred throwing it at 0-0 and 1-0 counts.






Location
This is probably the graph that needs the least amount of explanation. He throws inside and down toward the righties. This graph is not separated between righties and lefties, but it appears he never go low and outside to righties or low and inside on lefties. It would be nice if I had a better source for this data. Oh well.









Pitching Counts
A big knock on Olson last year was his inability to stay ahead in the count. I decided to break down his counts from last year and this year. First strikes were recorded 50% of the time (0-0 balls in play not included). This was up from last year when 44% were first strikes. To evaluate his progression through the counts, I created a graph. I did not include 0-0 counts in the calculations. Counts I included in the behind category were: 2-0, 3-0, and 3-1. These are situations where less pitches are required to result in a walk than a strikeout. I thought it more important to define counts on the endpoint as opposed to the progression. Even counts were: 1-0, 2-1, and 3-2. Ahead counts were: 0-1, 0-2, 1-1, 1-2, and 2-2. The result is the bar chart to the right. He is actually getting behind more in the count than he did in 2007. This should be a concern. Of course, it may just be a single game event. I'm not sure. If this was his "issue" last year, then it seems like he has done little to correct it.

Hit Quality
Another part of the puzzle may be that somehow he is able to induce poorly hit balls. This may be a skill or it may just be luck. Here is a comparison of important hit categories:
Category........2007..........2008
LD%..............18%............6%
GB%..............34%...........53%
LOB%.............66%...........78%
Ok, so what can you tell me from that? If we think what Olson did on Wednesday will continue, then it seems we think he is going to keep these new rates. Well . . . there is no way he can keep a 5.9% line drive rate. His batting average against (BAA) would be about .180 and his stuff is not that good. Now, has he become a groundball pitcher? That is a large shift and I am inclined not to believe that is one. I think when the rates shift to something more reasonable, the number left on base will also decline. I am completely and totally confident that he cannot sustain that level of batted ball success. It has been mentioned that he is hiding the ball better this year and if that has anything to do with it . . . David Blaine is a chump. Olson would be the master illusionist.

Conclusion
Having only looked at the stats and not one second of video . . . take this for whatever you think it is worth. Olson is young and is bound to improve, but his rates suggest he just might get shelled in his next start. I hope I am wrong. One game is not much to base an opinion on. He needs to improve on the number of walks he gives up, his counts, and somehow retain a high enough groundball percentage (or else those eventual long balls are going to knock him out of games). He definitely has talent, but I doubt he has arrived. If we are able to acquire some of his pitching performance . . . Stotle will analyze it, put it up here, and tell me that I am wrong.

01 May 2008

Breaking Them Down: Daniel Cabrera (Part 2 of 2)

To recap our findings in Part 1, Cabrera's April 2 start was plagued with inconsistent motion at the end of his delivery. This seemed to cause a loss of command -- primarily up in the zone. We now turn to his April 23 start, the third of three very effective outings in which he went 20.2 IP allowing just 5 walks and 16 hits. Our supposition was that Cabrera's delivery would be a little more uniform in this start, leading to better command and fewer walks and hits (due to an increased ability to keep the ball down). Let's see how it went.


April 23





Hmmmm. Unfortunately, it looks like Cabrera hasn't slayed his mechanical demons just yet. While there is still a lot to be satisfied with, the conclusion of his weight transfer is still inconsistent, meaning any sort of long term success -- at least to the tune of his last three starts -- will be difficult. One positive to take away is that he seems to be gradually moving towards a comfort zone with his follow-through, which may indicate progress towards a more consistent motion down the road. Let's look at the start in a little more detail.

As with the April 2 start, the middle-innings seemed to be the trouble spot. If you look at the 5th inning of the April 23 game, you see the follow-through gets a little more exaggerated, at times mimicking a full step towards 1st. Ideally, Cabrera will sit in his motion where he does at 0:10 in the above clip. His momentum is primarily towards home, though he still finishes with a soft fall-off to the 1st base side (if you recall, we found his motion to be at its cleanest in the April 2 start when the right foot finished between the plant foot and home plate). So which motion should Cabrera focus on moving forward?

As opposed to the April 2 start, it looks like Cabrera has begun to find a sort of comfort zone when his finish falls-off a bit to the 1st base side of the mound (wrapping his hip and leg across his body). The key to success, should he keep with this as his target follow-through, will be keeping his momentum towards home as long as possible. The right leg will act as a counter-balance at the release point and should be moving forward towards home. As soon as the release point has passed, he can fall off softly to the 1st base side by continuing to "wrap" his leg, which as mentioned above seems to be where he is most comfortable right now. If Daniel can focus on hitting that motion on a pitch-by-pitch basis, he'll put himself in the best position to consistently hit his spots.

We'll revist D-Cab later in the season to check his progress. Hopefully, Crawdaddy can add some color with a look at the f/x data, as well.

*This was written prior to this week's White Sox game, in which Daniel regressed a bit in walking 7 hitters on a cold and dreary Chicago day. Perhaps we can write this one up to the weather, but let's all pay attention to his follow-through when he takes the mound this weekend.

30 April 2008

Is our defense saving our pitching?


Before this becomes a Terry Crowley fan site, I wanted to delve into how well our defense is performing. Of course, to Oriole fans, defense immediately brings to mind the exploits of Brooks Robinson. Here on this blog, I think of Luis Hernandez. In the offseason and spring training, it was emphasized that one of the major keys to this team was going to be defense in order to help our fledgling starting rotation grow. Luis Hernandez was handed the keys to SS from the much criticized Miguel Tejada (somewhat unfairly, I still think). Luis has never mastered any level of pitching. Some could say his progression through the minors was largely an issue of social promotion. Now, Dave Trembley did some shootin' from the hip and declared that Luis has twelve days to show his stuff or will, presumably, be designated. We'll see. Anyway, the rest of the personnel movements were replacing same with same. Corey Patterson, a fleet defensive CF, was replaced by Adam Jones, a fleet defensive CF. Jay Payton, a defensive LF who is really neither, was replaced by former platoon player Luke Scott.

Null
Defense is neither saving runs or giving up runs.

Method
I decided to use RZR as it calculates efficiency in terms of ball entering the player's defensive zone. For each position (except catcher, which is not included in these calculations) I used all qualifying American League players and determined a median of performance. I think took the difference between the Oriole fielder's RZR and the median AL RZR. This percentage difference was then multiplied by the number of balls in the Oriole player's zone. The resulting number is the number of plays the player has made above or below median. This number was then multiplied by 0.600 runs as that is roughly the amount of probable runs when comparing 1 out with no one on and 0 out with a man on first. So, in other words, the result of not making or making a play was considered the difference between a single and an out. Outfielders were given a little extra help for earning assists. A typical assist would be worth about 1 run (difference between a double and an out). To be conservative, I put the worth at 0.75 runs and assumed all players were equally efficient at throwing men out. Finally, the runs saved or given were added up to produce a total runs saved or given. This total was then divided by 10 to determine number of wins earned or lost.

Results
Mora has been the most efficient position player based on efficiency and opportunity. He has saved 2.46 runs (+4 plays). Jones' fielding is also above average with 0.95 runs saved (+1.5 plays). Scott and Markakis both cost the team runs with their fielding, but make up for it with their assists. They save the team 0.42 and 1.97 runs, respectively. Fielding-wise Scott is at -0.5 plays and Markakis is at -1.75 plays.

The other positions have cost runs. The most egregious offender is Mr. Defense himself, Luis Hernandez. He has cost the team 1.44 runs, missing 2.5 plays. Next worse is Kevin Millar with 0.53 runs given (-1 play). Roberts is almost the median player. He costs the team 0.13 runs (-0.25 plays).

When we total that up together (ignoring the backups who have logged time), we come to 3.69 runs saved. That comes to 0.369 wins earned based on defense. If we keep this level of play for the entire year . . . we will have earned 2.2 wins.

How much can we gain by replacing Luis Hernandez mean?
Well, let us assume that the next player is just as awful hitting the ball. This will be a conservative estimate. We'll assume that whoever replaces Luis has league median UZR. That would be worth an extra 0.9 wins to yield 3.1 wins. Add that to the probable fact that they are most likely worth a win or two more than Hernandez with the bat and it is conceivable the team would be 5 wins better with someone else at shortstop.

I'll take a look in a few days if such a ball player exists in the Orioles' system.

29 April 2008

Keep On Keepin' On

Well, we now have 5 data points, so I think it is now a good time to convert this over to a chart format. The chart details the total season wins predicted by PECOTA and ELO for each week. After every 40 games, I will readjust the ZiPS/Morong prediction that I arrived at. Finally, I am including the actual number of wins.

The prediction models now has as a 100:1 (PECOTA) and 8:1 (ELO) odds of making the playoffs. These are our best odds of the year. The two major questions we have as we stand at 14-11: 1) how representative are the player's performance in the first 25 games for the entire year? and 2) how representative has our schedule been in relation to the rest of the year?

28 April 2008

Breaking Them Down: Daniel Cabrera (Part 1 of 2)

If you're an Orioles fan, you've come to find lines like this familiar over the past few seasons:
Game A
5.0 IP, 2 H, 3 ER, 7 BB, 4 SO

Game B
5.0 IP, 4 H, 4 ER, 7 BB, 6 SO

Can you guess the starting pitcher? Okay, it's a trick question. These were actually lines posted this past Thursday night by Tom Gorzelanny and Dustin McGowan, respectively. Of course, no one would hold it against you if Daniel Cabrera was the hurler that came to mind. In fact, Daniel started the season much like the Daniel Cabrera of old:

April 2 vs. TAM
4.0 IP, 6 H, 6 ER, 5 BB, 2 SO

April 7 vs SEA
6.0 IP, 5 H, 4 ER, 4 BB, 5 SO

Since then, however, Cabrera has strung together three straight solid outings, with an impressive combined line:

20.2 IP, 16 H, 5 ER, 5 BB, 12 SO

That is an incredible 2.23 ERA and 1.04 WHIP for the normally combustible D-Cab. To quote MJ, a good friend and co-owner of a fantasy team with both McGowan and Gorzelanny, "The only pitcher not pitching like Daniel Cabrera is Daniel Cabrera!" So what has been the secret? This 2 part piece breaks down two starts -- April 2 and April 23 -- taking a look at Daniel's mechanics to see if there are any clues as to the reason behind his apparent turnaround.

APRIL 2
Let's start by looking at April 2, a start in which Cabrera struggled to keep runners off base, allowing 6 hits and 5 walks in 4 innings.




Two things jump out. First, Cabrera was very consistent with his release. He's finding the same arm slot and release point in each of the pitches shown. Second, periodically he tends to slip into a little extra movement at the end of his motion -- essentially forcing his follow-through on his back leg all the way across his plant foot and to the left side of the mound. Let's take a look at each of these one at a time and talk about their significance.

One of the most difficult tasks for a tall pitcher is to find a consistent arm slot and release point. Often times, younger pitches will struggle to achieve consistency in this area, which is a huge reason why a "repeatable" delivery is one of the more important characteristics a pitcher can show a scout. Cabrera seems to be much more consistent this year in his motion, and as a result is finding a consistent arm slot and release point. Curiously, as you'll note in the video, the resulting pitches are not necessarily improved. That brings us to point number 2.

Ideally, a pitcher wants his motion focused towards home plate once he starts toward the catcher. Cabrera isn't bad in this regard. At the end of his release, however, his right leg will swing around and land anywhere from directly in front of his plant foot (left foot) to two to three feet to the left of his plant foot. These inconsistencies translate to erratic pitch location -- namely, in Cabrera's case, elevation in the strike zone. The 1st follow-through is generally less exaggerated than the 3rd inning follow-through, and likewise the 5th. Cabrera seemed a bit out of sync towards the end of his motion, and simply lost command as a result. This is still a large improvement from the more serious issues D-Cab has had with his mechanics in the past, but one would expect it to be difficult to find consistent results if Daniel can't work to conform the last bit of weight transfer towards home. It looks like his ideal motion drops his right foot almost directly in between the left foot and home plate (see 0:27 and compare with the two pitches following).

So, you might guess correcting this is the key to his success over the last three starts? Well, take a look in Part 2...

Eating Crow 2.2: Murray and Ripken


Before launching into this exercise, I'll recap the previous Crowley entries. In the first one we evaluated his patience and contact rate as a player. We found that his contact rate was higher than league average, but his patience was league average. In the second study, we tried to discern whether changes in pitching coaches altered patience and contact rate on the Orioles. It did appear that batting approach did change slightly toward being less patient during Crowley's stay. It also appeared that the talent level may not have been high enough to make that approach work as contact rate was somewhat below average.

A conservative way to go about determining if a hitting coach affects the hitting approach of his team may be to focus on the team's star players. It follows reason that star players are those who are the most likely to retain their previous approach because, in their mind, it is what made them a star in the first place. I would assume that these established players would probably not pay much attention to what their hitting coach would say. This is doubly true as both Murray and Ripken, the Orioles established stars, played with Crowley.

Null
The hitting coach does not affect plate patience in established stars (Eddie Murray and Cal Ripken Jr.).

Method
As used in the previous two exercises, we will use estimated pitches per plate appearance to measure plate patience. Contact rate will also be measured as it is a type of skill that works well for an aggressive approach. Murray and Ripken will be compared individually, to the team, and to MLB. The time frame we will use is from 1981 to 1991. This includes four seasons coached by Rowe, four by Crowley, and three by McCraw (except for Murray who was traded to the Dodgers for those final three seasons).

Eddie Murray
Eddie Murray was perhaps the best 1B in the game during this stretch. From 1981 to 1984, he was clearly the best 1B in MLB and perhaps the best player. He was never a prolific homerun hitter, but he had plus power, plus contact, and good patience at the plate. During this time, he scored 156OPS+ every year during this stretch. That is consistency. Crowley was hired for the '85 season and Murray's performance eroded. Murray was healthy and should have been in his prime (Age 29), but he may have just peaked early. His next four years were 149, 136, 120, and 136; which is still quite good. It is interesting to see his strikeouts and walks decreasing along with power. Typically, when walks and power decreases . . . strikeouts increase as pitchers are testing the batter. It is a sign of reduced bat speed. He would never regain that tendency to walk as he had during that amazing four year stretch.

Cal Ripken
Cal Ripken seems like a guy who has little faith in his hitting coaches or way too much. Going to a game was often to provide you with a myriad of batting stances. One thing I am not too sure of is whether this was true very early in his career. He second and third best offensive seasons were 1983 and 1984. He won the MVP in '83 and should have won it in '84. His best season came in '91 where he once again won the MVP. Cal was not like Eddie. He has always seemed to be a bit more aggressive at the plate. Anyway,you all should be well aware of his exploits, so I'll move on to the results.

Results

Pitches per Plate Appearance
Murray's numbers were above average in terms of plate patience prior to Crowley's arrival. After his arrival, Murray's patience was league average for the entirety of his remaining years with the Orioles. Cal's patience was somewhat below average, but decreased even further during his first two years under Crowley. His final two years under him shows a marked increase in patience. This trend would not continue into the McCraw years. It is interesting how plate patience decreased once Crowley began coaching. It also seems that in '87 and '88 Cal began taking a different approach. I can't really blame talent level on Cal's progression as he had Eddie performing above average behind him. I doubt people were pitching around Cal. Also, Cal typically did not come up with anyone on base. This is especially true in '88 when Frank Robinson was all hopped up on pills and let Billy "Gold Glove, Lead Bat" Ripken take in 540 PA, primarily, in the 2 hole. That is 207/260/258 for the year and we think Luis Hernandez is bad (sorry, that is another topic). Anyway, it looks like Cal changed his approach in '85 and '86 to be more aggressive.

Contact Rate

I at least don't see much of interest here. You can tell when guys are patient and when their skill levels increase. They are both players who could do well with an aggressive approach as they both do a good job with making contact. For Eddie, it looks like a change in his approach. For Cal, it looks like his growth as a hitter.

Conclusion
Well, like last time, the data is somewhat confusing, but it appears that Crowley's philosophy about aggressive hitting was embraced by Eddie all four years and by Cal for two years. I think that is more definite about Cal. Eddie may have just seen a regression in his abilities, but losing the ability to take a walk so soon after establishing it as a hitting trait seems unusual. That is why I think this data indicates that they listened to him. So, if established stars are listening to Crowley . . . I imagine the rookies and those desperate to stick on with the team are also listening to him. I think players probably stick with their approach until they think they are in a slump and then seek out advice. A hitting coach is supposed to know what he is doing, so they listen. I think Crowley's suggestions may work for some, but I don't know if we have found who yet.

24 April 2008

Eating Crow 2.1: Crowley's Affect on the Orioles '85-'88


Terry Crowley was the hitting coach during the dark stretch between the '83 World Series win and the '89 Why Not? season (of course the answer to that question was because after Bob Milacki and Jeff Ballard . . . we had awful, absolutely dreadful starting pitching). Anyway, Crowley, fresh out of retirement, plied his trade and tried to mold Oriole hitters. If he has influenced his teams, we should see a decrease in plate patience measured as pitches per plate appearance. We may also see an increase in contact rate as he teaches hitters to actively pursue pitches on the fringe of the zone, which are often poorly hit.

The null here would be:
There is no difference between Crowley-influenced Orioles and non-Crowley-influenced Orioles.




Methods

This section will be handled simply. Crowley's team hitting will be compared to Ralph Rowe's teams (81-84) and Tom McCraw's (89-91). The measures will be P/PA and contact rate. These measures are explained more in depth in the first entry on Terry Crowley. This study will assume that the player's are relatively uniform in quality and openness to be taught.

Results

Pitches Per Plate Appearance
Ralph Rowe's years with the Orioles are notable in that his teams we slightly above the league average in pitches per plate appearance. The differences are not great, but are significantly above the league average (Rowe = 3.71 +/- 0.02; MLB = 3.67 +/- 0.01). Crowley's team only had one year noticeably above the MLB average and it was his first season. I wonder to what extent Rowe's hitting concepts were still being applied as it was Crowley's first year coaching and hitters typically stick with what works until they struggle. It appears as if, under Crowley's watch, the hitters are getting less patient. It is also interesting to note that in 1986 and 1987, there is a weird blip in terms of P/PA. I have no explanation. I know 1987 was the year in which an uncharacteristic number of homeruns were hit. Finally, McCraw's first and second year were typified by a major correction in P/PA. His third and final year had the team back at league average. The main explanation here would be Mickey Tettleton leaving, Worthington replaced by Gomez, and Ripken changing his approach. It appears McCraw also professed being patient. I am actually quite surprised by what this data seems to say.

Contact Rate
The metric seems to relate well with P/PA. They should be inversely related to each other. As plate patience decreases, contact rate should increase. This is due to coaching encouraging the utilization of balls located on the boundary of the strike zone. Likewise, batted balls are more likely to become productive hits as plate patience increases. I have a feeling Rowe's results that are above average might be due to the team being more talented offensively during his strand. During Crowley's tenure, it is a bit all over the place. McCraw's time is characterized by below average contact rate. This all makes sense with what the data seems to be suggesting, but the data here is hard to parse.

Conclusion
These are in no way conclusive results, but strangely . . . it seems the Crowley might have an effect on team batting behavior. It appears, maybe superficially, that his teams are less patient than Rowe's or McCraw's and that his hitters go after pitches that may not best utilize their at bats. It seems that the head coach does not seem to affect these metrics as Weaver/Rowe did not have the same result as Weaver/Crowley. Although, Rowe's teams were superior and this may be affecting the data. I'm not sure how to normalize talent level.

Next Time
I am going to still keep at Crowley's first tenure with the Orioles. I will analyze his affect on hitters at different career points and see if there are any tendencies.

23 April 2008

Eating Crow: A Tendency for Aggression?

Terry Crowley is a name many Oriole fans despise. They complain about how he supposedly affects Oriole hitters. He preaches being aggressive at the plate. Others, myself included, think that hitting coaches do quite little at the Major League level beyond offering useless advice and acting as a guidance counselor . . . as in someone who is good to talk to, but doesn't really change things. In this piece I will try to determine if Crowley has a history of affecting his team's offensive production. This project will consist of four parts: Crowley as a player and Crowley's three stints as a MLB hitting coach.

Our working null hypothesis is that there is no difference between Crowley's performance and his team's performance against league performance as measured by pitch counts, a measurement of aggressiveness.

Part I: The Player

Terry Crowley was drafted by the Orioles in the 11th round of the 1966 amateur draft. He progressed quickly through the Orioles farm system: 154 games at HiA Miami, 55 at AA Elmira, and 2007 at AAA Rochester. He really didn't show much statistically until his second go around at Rochester in 69 where he displayed plus power. Once he reached the majors, he showed that he just was pretty average as a hitter. He never developed into anything more than a pinch hitter.

Methods
First we need to convert his statistical line to pitches per plate appearance. I do not have that data and I doubt they kept track of pitch counts by batter back then, so we'll have to estimate this. The basic formula is as follows:
[3.3*(plate appearances) + 1.5*(SO) + 2.2*(BB) ] / (plate appearances)
We will then compare his P/PA against the league average. This should be a decent indication of his aggressiveness as a player. It should be noted that we are dealing with a small sample size. The years with more than 100 AB are 70, 72-74, 78, and 80-82. Other metrics will also be compared: contact rate. Contact rate will relate to success of aggressiveness.

Results
Based on the chart, Crowley was not a hacker. He was about league average. If you ignore the seasons where he had less than 100 PA, He excedes the league average his second year and toward the tail end of his career. During , his prime . . . league average. This somewhat backs up Bill James' often cited, but never conclusively evidenced theory of age and walk rates. Basically, it goes that as players realize their bat speed is gone, they compensate by taking more pitches and waiting for the right pitches to hit. The result is that a player will start walking at a higher rate until the opposition realizes that he can no longer hit. Then he will bottom out completely. That describes Crowley's tail end Oriole career and horrendous stint with the Les Expos.

Interestingly, Crowley's contact rate is predominantly better than league average. It typically decreases as the more patient he is, but not every year. His only years with over 100 PA, his contact rate was much better than average. Here is where the issue of his theory of aggressiveness may lie. His patience is somewhat fluctuating, but his contact rate is always exceptional during the seasons when he was given significant at bats. Even in his final year, he displays a very good contact rate, but he was unable to do anything with it.

Conclusion
So, let's revisit our null. Is there any difference between Crowley's and the league's P/PA? Crowley comes in with a 3.72 +/- 0.11 P/PA. The league comes in at 3.69 +/- 0.02 P/PA. Crowley's numbers are NOT different from the league over the course of his career and there seems to be no general pattern except that he seems more patient during the beginning and end of his career. When contact rate comes into play . . . things get a little hazier. The seasons in which he was rewarded with significant at bats were those that also we years where he was successfully aggressive. It seems as if, as a player, his approach was never based on plate patience, but looking for a pitch he could make contact with. Contact rates over the career also do not differ from league contact rates, but it appears that Crowley had focused on simply making contact.

Contrary to the current perception of Crowley, he was not an overly aggressive hitter, even though he had a notoriously violent swing. He was selective, but not overly aggressive. He was not a hacker. As we know from ex-players like Joe Morgan and Billy Beane, a lot of old ballplayers think quite differently about the game now than what they did back then. Joe Morgan is notorious for his stubbornly held views about run production and Billy Beane reached epiphany, realizing why he was such a horrible talent. It may be that Crowley has bought into the perception of himself being the free swinging buck who was very much aggressive at the plate. Why he would teach this to others? I have no clue as he was a pretty average ballplayer with a 104 lifetime OPS+.

Next Time
The next segment will analyze Crowley's affect during his first stint with the Orioles. This will include several sub-studies. The first one will be discerning whether the team as a whole was more or less aggressive than MLB as a whole. After that we will compare his affect on inexperienced players and experienced players as well as free agents. It will be interesting to see if he has designed his team to be selective, but target contact.