07 July 2016

Orioles in First, Which is What Our System Projected

Alleged security camera recording of an alleged drunken,
shirtless alleged Steve alleged Melewski allegedly
destroying alleged computer equipment at an
alleged Best Buy on allegedly York Road, allegedly.
Back in March, mouths were foaming about how the projection models rated the Baltimore Orioles.  Steve Melewski was alleged going to local Best Buys, drinking a six pack in the parking lot, and then going inside to jeer at salespeople in the tablet section.  It was a dark time in Baltimore, but a rather proliferous one for message boards and writers who needed inspiration for fan friendly columns.

Truth be told, the projection models were not exactly in unison.  While Baseball Prospectus' PECOTA had the Orioles at 72 wins, ZiPS and FanGraphs pegged the club at 81 and 82 wins, respectively.  Some took those rosier projection harshly as well without acknowledging that the first place clubs were sitting with 84 wins in those projections.  In other words, two model approaches saw the AL East largely flattened while PECOTA was off on its own doing its own thing, which is what more recent iterations of PECOTA has been known to do as it has been more aggressively updated in the past few years.

At the Depot, I usually put together projections, but time was rushed and I simply was more focused on my CRAP draft system than updating my BUCK patch model to the ZiPS system.  In my gut, the club felt like a fourth place team around 82 wins in a pretty competitive AL East.  Other clubs looked better like Toronto and Boston, but the difference was not all that great.

In the first few weeks of the season, the Orioles won and won often, but typically not all that pretty.  The rotation was a mess, but the offense was humming and the club got some good fortune.  It meant that the data analyst cautious approach continued to note that the Orioles looked uneven and ready to fail.  Keith Law noted that without the winning streak at the beginning of the year, the club had played .500 ball.  Yes, we all know that removing a winning streak makes a team look worse and that there is typically no good reason to ignore a week and a half of games.  Anyway, an allegedly drunken Steve Melewski allegedly ran wild in several more local Best Buys.

In June, I decided to finally go back and figure out what my Buck patch model would have projected in the AL East.  I updated the system to consider the past three seasons each manager managed and each General Manager general managed.  I looked at what ZiPS thought they would do and then what they wound up doing in order to get an idea of the impact of those positions on the success of a baseball club.  All of this is based on my previous work, which I still have trouble believing.  For those without much experience, I simply projected them to not deviate from ZiPS in those seasons.

AL East
adjCD
RankCD
Rank
Baltimore
85
1
1
New York
83
2
4
Toronto
82
3
2
Boston
79
4
3
Tampa Bay
77
5
5
 
 
 
 
AL Central
 
 
 
Royals
92
1
2
Cleveland
88
2
1
Chicago
79
4
4
Detroit
77
5
3
Twins
80
3
5
 
 
 
 
AL West
 
 
 
Houston
88
1
2
Texas
84
2
1
LAAA
81
3
5
Mariners
80
4
3
Oakland
76
5
4

The patch model elevated the Orioles over the opposition as it considers the work by Showalter and Duquette to be exceptional with respect to expectations.  Over 2/3 of the projected ranks are off no more than one from where the clubs currently sit.  The biggest misses are the Twins who have been simply gruesome in part due to a heavy reliance on Minor League data for their projections.  The other big miss is the Rangers who have been blowing up other clubs this year even though they are a bit saddled with an aging and sleep deprived Prince Fielder.

All in all, it seems the data analysis approach has worked remarkably well for most of us doing this.

7 comments:

Roger said...

From what I've seen the O's are outplaying their Pythagorean Win projection by about three wins, which is at least the difference between first and third in the division. The AL East is currently the most competitive division. Also, all the models were cool on Kim, Trumbo, and Schoop (Fangraphs still doesn't like Schoop) who have all performed beyond expectations. Also, the last discussion we had about unearned runs certainly tainted Fangraphs' projections. There are pieces in the pitching staff that have outperformed expectations too. Tillman, Bundy, Worley, Brach, and Wilson (yes, Wilson) have been better than expected too (I still like Wilson over Wright). So, while the starting pitching is pretty dreadful (especially at Safeco, oy), it has outperformed expectations. Obviously Matusz, Despaigne, Duensing, Miranda, and Aquino (although we know little about the latter) are or were not answers. The O's need a SP and a LOOGY that are reliable and that should be enough to put them over the top. I hope the rest can keep up their performances, too.

I am still kinda gloating a little that my idea of dumping Miguel and Matusz and that Trumbo was a genius pickup even with his defensive liabilities seems to have been the way the O's ultimately went. The O's should go after Brad Hand as a LOOGY and future starter. Would be a great value pickup. Aquino and Donnie Hart might make good LOOGYs too. Bundy is looking more and more like a viable starter for later in the year. He was smokin' hot yesterday at Dodger Stadium (96-98mph).

Jon Shepherd said...

First, lets not conflate the Pythagorean Win EXPECTATION with these other projection models. Completely different. PWE is a retrospective outlook and performs about as well as general record for predicting future performance (as would be expected because they track with each other in design). Second, a difference of three wins is well within the noise for PWE and actual performance.

Second, comparing actual performance to the projections is a bit off because we are comparing a mixture of outcomes to the 50th percentile. The issue is more about comparing teams on the same level. But, yes, the models did slightly underpredict performances for most of the Orioles. I think Kim's performance is the only exceptionally great one in contrast to his projected performance which was as a solid regular. I am unsure what you are trying to say beyond that as a club they are outperforming their projection, which is true for all first place and second place clubs as would be entirely expected.

Third, Fangraphs issue with unearned runs would have maybe, maybe, benefited the Orioles by two games.

Fourth, I think paying 1 MM to get a longer look at MiGo was smart based on the cost of starting pitcher. Unfortunately, MiGo would be pretty valuable for the Orioles and would be second or third in ERA if on the team. That said, I wanted to get rid of MiGo as I wrote about in October. I think the Orioles' solution was better than mine. I also wanted to deep six Matusz, but I can understand his ability as a LOOGY was important to the club and they have had difficulty getting anything close to what he did in the past. We were all largely in line with the Trumbo pickup. We thought it was a good deal. Yes, Hand looks good to you, the Orioles, and several other clubs with best prospect portfolios than the Orioles.

Fifth, I would drop about 3-4 mph off Bundy in a starter's role as was generally seen before all the health injuries. It is good to see him effectively mix three pitches, which I think is far more important than his velocity. I would hesitate having him start this year. I would seek out all alternatives before that. If the club has run away with it in September, I would be open to him having two or three starts on extended rest. As you can see, he is gunning it. He would need a different approach as a starter.

Roger said...

I agree with your concept of equal comparison at the 50th percentile. But we can't just say that everyone outperforming their projection is just doing that without thinking that some of the projections may have been wrong. It appears that BUCK did a better job then many of the others. I do think the projections were somewhat off specifically for guys like Schoop, Tillman, and Trumbo. I looked at them and instantly thought they were too low. These guys fall into two categories - positive regression and progression. I think it was fairly apparent the Schoop was a breakout (progression) candidate and the projections should have reflected the possibility better. I think it was also reasonably apparent that both Tillman and Trumbo were candidates for positive regression to their prior performance level. Tillman showed flashes last Fall and Trumbo had a good second half. These three alone might have added a couple of wins for the O's in the projections.

As far as Bundy goes, yes, I agree. I think Buck has him on a 60 pitch count. Does it matter if these pitches are thrown at the beginning of a game or in the middle? If he can give 3-4 solid innings as a starter every five days, how is that not better than Ubaldo (or Wilson or Wright)??? Give the O's hitters four innings to get a lead then hand it over to the bullpen.

As far as Hand goes, yes, he's pitching great and, yes, he was once a top prospect, but the Marlins gave up on him and the Padres just picked him up off waivers. If most clubs don't believe he can eventually start then there may be a chance for the O's to make a deal. I am an optimist. Another LOOGY candidate that might be had cheap (and I can't believe I'm saying this) is Ian Krol, who's turning in career best numbers for the Braves.

Jon Shepherd said...

If I flip a coin and it comes up heads 20 times and tails five times, it does not mean that we have to think that the coin is not balanced. There is a great chance that those results will happen by chance. If we are convinced of something in the midst altering those projections then we should be putting forth an actual mechanism why things are off. However, I am at a loss to find such a mechanism as the players simply do not all fit in one bag beyond the power component (which is not universal as Kim is not in that group). This is the year of the home run and a dependency on past data for projection would perhaps miss such a large deviation in performance. That might explain a few if it actually does explain those few. It would not explain all of them.

I think your view about Tillman and Trumbo are decent candidates for improvement, but not based on your suggested evidence. Trumbo would be a season long decent hitting while playing in parks that were not well suited for his ability. A bounce back was expected and suggested by ZiPS with some restraint. Tillman was good for a while, had a tough season, and we bound to improve. Schoop was considered a breakout candidate, but a 50th percentile projection will not show that. You need a full range of outcome projection to do that. I am unsure what you are trying to suggest here.

Does it matter where Bundy pitches? Maybe. Some pitchers desire greatly defined roles. Startign without expectation of 5 or more innings might alter approaches or mentality.

Padres were one of the first teams able to select Hand with waiver wire priority.

Roger said...

That's true about the Padres and Hand. I know. Just trying to have some optimism. Let's have a blog post to discuss who the O's should pick up and why. I know we had the Nolasco post but that one seemed weak.

Where Bundy pitches could easily matter if you think that someone like Wright or Wilson is more suited to delivering in the middle innings (giving them, say, a two inning limit as bullpen pieces). Both have shown the ability to do that. Whereas, if you have Bundy on a pitch count limit, you don't know how long he'll last. Plus you start getting him ready for next year.

On projections, the point is (and I think you partially made it) that last years statistics aren't good enough to project this year's numbers. There has to be some qualitative analysis as well as quantitative analysis. I know that statistics driven gurus flinch at the idea of anything they can't measure but sometimes your eyes tell you things that statistics can't. I'm saying that the predictions for those three players were "predictably" low and that they didn't reflect what the true 50th percentile should have been for them. To turnaround your metaphor in the case of the coin, the flips you mention are likely random because you can measure the coin to verify it's neutrality. When you're talking about human performance, it's not so easy to verify the subject's neutrality to projection so it makes sense to take a step back and re-evaluate when the subject performs different from what's expected. It doesn't matter (much) if a coin is 26 or 126 years old when you flip it but it does in humans. Schoop is young enough to not have finished developing his adult body. Also, between the 25th and 50th flips, a coin is not likely to gain or lose weight, but a human will (or muscle mass). These qualitative differences could make a player's 60th percentile by past statistics into his 50th percentile going forward.

Jon Shepherd said...

With projections, it is taking several years of data and then utilizing all applicable historical data, so it is a bit more than you suggest.

Second, qualitative analysis is a major part about how to utilize data in general. It is not something that scares analysts. Analysts love data whether it be quantitative or qualitative. Analysts dislike hunches.

I think my only point is that we do not have the data to suggest there is something inherently wrong with the projections. They do fine with 50th percentiles. There were reasons to think Trumbo or Schoop would do well, but them doing well does not mean the projections were off.

I think Matt K is putting something together on pitching.

Pip said...

Dumping a solid player with remaining options just to save 2 million dollars in salary is very foolish and ranks among the Dumbest in a long line of dumb moves by Duquette.