05 May 2016

How To Fix The Orioles' Biggest Flaw

The Orioles have a huge problem. Actually, they’ve got a number of them. I mean, it isn’t good that none of Rickard, Kim or Trumbo can play passable defense in the outfield and therefore the Orioles have two competent defensive outfielders. It is worrisome that the Orioles have no infield depth to speak of, and JJ Hardy will be out for the next six weeks. It’s problematic that the Orioles basically have four 1B/DHs on the roster: Pedro Alvarez, Mark Trumbo, Hyun Soo Kim and Chris Davis, and have a fifth in Jimmy Paredes expected to return from the DL in the near future. It can’t be good that Zach Britton is day-to-day, that Dylan Bundy has two strikeouts in seven innings, and Hunter Harvey is back on the DL with yet another injury.  But all of that can be worked around and pales in comparison to their real problem. The real problem is their starting rotation has been a dumpster fire or a complete disaster.

At the team level, it doesn’t look so bad. Sure, they have a 4.62 ERA good for sixth worst in the AL. And yes, opposing batters only have a .218/.336/.294 line with runners in scoring position with zero home runs, a .260 BABIP, a 13.5 K% rate and a 10.1 BB% rate. Opposing batters are going to start hitting for power with men on base, with potentially disastrous results for this staff.  It isn’t a good sign that the Orioles rotation is averaging only 5.29 innings per start, which is the eleventh lowest rate in the majors since 1974 (total sample size is 1294). In the interest of full disclosure, the Reds (10-16), Brewers (10-15) and Rockies (12-13) starting rotations all average fewer innings per start than the Orioles. And let’s not forget the Orioles only have 8 quality starts so far this season, good for a 32% rate. League average is 48%.

But if you ignore those small details, they also have a respectable 3.97 FIP and a 4.14 xFIP.  They have a surprisingly poor LOB% of 68.7%, good for third worst in the AL. It seems possible that some of the starting rotation’s problems should be blamed on the defense and that the rotation is middle of the pack. After all, why would a defense with an Rickard, Jones, Trumbo outfield and having JJ Hardy on the shelf struggle?

But the problem comes into focus when you look at the pitchers individually. Chris Tillman and Kevin Gausman have done well. Tillman may only be averaging 5.33 innings a start (although a severe rain delay hurt him in this regard), but he has a 2.81 ERA, a 26% K-rate and has been worth 1 fWAR.  Aside from one game where he struggled, he’s been excellent. Gausman has played in only two games, but has a 2.45 ERA, a 22.7% K-Rate and a 3.53 FIP.  All in all, these guys are showing ability.

But the rest of the rotation has been terrible. Tyler Wilson is the only other starter with an ERA below 5.00 as a starter, and has an FIP of 5.35. He’s extremely hittable with a strikeout rate of just 12.2% and a walk rate of 2.4%. He could possibly be serviceable with a good defense, but the defense has some issues. No, I don’t care how he has performed as a reliever. Vance Worley has an ERA of 5.06 and an FIP of 4.57. He’s got the best K-BB% on the team at 17% and has an excellent K% rate of 23.4%. He’s struggled due to a high HR/9 rate and a low LOB%. Ubaldo Jimenez has been his usual Jekyll-and-Hyde self with two good starts and two poor ones. He has an ERA of 5.20, but has an FIP of 4.70 and an xFIP of 3.92. He’s likely better then he’s performed so far, although it isn’t likely that he’ll be a staff ace. Gallardo and Wright have also struggled, and Gallardo is currently on the DL and throwing an 88 mph fastball when healthy. Jimenez and Gallardo are the only two guys in this list that have a track record suggesting that they can start, and Gallardo certainly has had issues to start this year. Unless you believe in Wandy Rodriguez, it’s not like the Orioles have much help available in the minors. All in all, it’s pretty clear that the Orioles only have three starting caliber pitchers.

The Orioles could look into trading for a starter, but without a farm system or excess payroll it’s hard to see who they can add. Also, it’s questionable how much one starting pitcher will help their rotation given that they need three. Not to mention that it’s just the beginning of May and it’s hard to trade for a pitcher before July.  The Orioles have roughly another sixty games until the all-star break and probably need a solution before that point of the season.

This rotation is going to make it hard for the club to win games and is going to kill the bullpen. Most of these starters largely can’t go deep into games and aren’t effective even in their limited playing time. That’s why the Orioles should go outside of the box and use a four man rotation.  But I have a different idea of how that would work than the usual.

Tillman and Gausman would be the #1 and #3 starters respectively. They pitch every five games. If you have two legit starters, you might as well use them. The guys in the #2 spot are Jimenez/Worley and the guys in the #4 spot are Gallardo/Bundy (with Gallardo hurt, Wilson takes his spot). The first starter goes through the order twice and the second starter goes through the order once. This results in the two starters facing 27 batters and hopefully going through at least six innings but hopefully seven. This allows for a six man bullpen of Britton, O’Day, Brach, Givens, Matusz and McFarland.

A team can play a maximum of twenty games without an off day. So, the schedule would look something like this:

As you can see, Tillman and Gausman are on a regular schedule. The combo starters get between three and four days of rest per start. This is normally considered short rest, but these pitchers will have a much reduced workload. In addition, it is possible to have Worley go four innings instead of Jimenez every once in a while and to have McFarland start instead of Gallardo, Wilson, Wright or whoever. Wilson, Wright and McFarland all have options remaining and therefore it would even be possible to swap them out regularly to ensure that each gets enough rest.

The starting pitchers wouldn’t be very happy with this plan. Gallardo and Jimenez aren’t going to be getting many wins if they can’t throw five innings. But Gallardo is throwing his fastball at 88 and has a 7 ERA. If he can’t improve on that, it’s not like he’s going to be in huge demand come next free agency. Jimenez is performing better, but he’d have a chance to factor in the decisions of over 40 games.  With decent performance, he could easily be a twenty game winner.

This plan would mean the bullpen is a man short. But it would also mean that the rotation would have a better chance of going six or more innings in a given night. If these pitchers can go deeper into games, then the bullpen will remain in better shape. This plan would require Tillman and Gausman to successfully produce. If they struggle, or aren’t able to average six innings an outing on a regular basis, then this will be problematic. But the alternative is hoping that starting pitchers that have been poor so far and have a bad track record will suddenly become good. I wouldn’t hold my breath.

The Orioles do have a lot of flaws, but they’re largely offset by the fact that they’re able to crush the ball. The offense, with a wRC+ of 119, is ranked the best in the AL. They’ve scored a respectable high 4.5 runs per game. Their hitting will make up for a lot of weaknesses. And if the Orioles are willing to go outside of the box to fix one of their largest flaws, they may just defy the projections yet again this season.


Roger said...

Good article and agree with the approach - makes a lot of sense. Halfway through, I was just thinking about the stat they posted during last night's game on how Wilson was extremely effective the first two times through the order and had an ERA well over 5 on the third pass which was certainly consistent with the actual performance last night. I'm not sure I would characterize this (if I were trying to convince Buck) as having only 6 relievers. I would characterize it more as having four starters, eight relievers, and using three of the relievers as swing men. What you're really saying is we have two good starters (Gausman and Tillman) and six spot starter/swingmen (Jimenez, Gallardo, Wilson, Wright, Worley, McFarland). Recognizing this and using it to the O's advantage could really boost the team. I honestly think the key to how you use them is to look at their records in the first inning. The guys that can make it though their first inning are better relievers. Starters tend to try and absorb a little damage in the first because they plan to pitch more innings total and want to pace themselves. To me that might make Jimenez, Gallardo, and McFarland the starters with Wilson, Worley, and Wright as the relievers. Planning to change pitchers BEFORE they get clobbered is certainly a good strategy.......

Matt Perez said...

The problem is that your plan would have five starters and seven relievers. But two of those relievers couldn't be used regularly because they'd need to be saved for long outings when the other three pitchers start. Which means you'd only have a five man bullpen and would have limited flexibility. What do you do if Gausman goes three innings one game? The beauty of a four starter system is that you have more flexibility.

It would also mean that Jimenez, Gallardo and McFarland would throw roughly 130 innings a year and would never factor into any wins. I don't think you want McFarland going that many innings and you probably want Jimenez and Gallardo going more than 130 (presuming health). Also, Jimenez and Gallardo would be unhappy about this. The plan I'm talking about allows Jimenez and Gallardo to pitch in roughly 30% of the Os games or at maximum in the 190-200 range. More likely, we're talking 170-180, but that's a full season.

The other thing is that Buck would probably think of such a plan as taking away a reliever because he couldn't use Bundy or Worley regularly. It's possible they'd be available every once in a while, but most likely they'd be pitching every four games. Such a situation would be ideal for someone like Bundy who needs innings but takes a long time to warm up.

Roger said...

Don't get me wrong, I'm not disagreeing with the four man rotation just pointing out what seems to be the best "roles" for the pitchers discussed. Not sure why you want Jimenez and Gallardo going more innings, though. Don't care about their wins that much - they're making enough money to pitch how they're told to pitch. Need to maximize TEAM performance not try to "give" wins to specific pitchers. Even with the success so far this season and the short outings from the starters, we have seen a helluva a lot of bad fifth and sixth innings. I'd rather have a Wilson with a 2.30 ERA and be saying "I wonder what he'd have done if we let him go five (or six)" instead of a Wilson with a 4.50 ERA saying "I sure wish we hadn't let him pitch the sixth".

Roger said...

Also, I'm not sure Worley/Bundy are the best choices for the "second" starter. I think Worley/Wilson might be better. Hold Wright and Bundy back for the times when Gausman/Tillman don't cut it. Bundy is likely to be the fifth starter at some point when he's had enough recovery time. And Wright is too inconsistent for a regular role.

Matt Perez said...

The thing about Worley/Bundy is that they're both out of options so they need to stay with the big league club. You can only have one of McFarland/Wright/Wilson on the roster (the other two get optioned) when everyone is healthy unless the Os go with an eight man bullpen as they are at the present. Or you have to jettison a player. Option logistics play a role in my plotting.

I also like the Bundy/Gallardo paring because it mixes different velocity levels and also lets Bundy pitch on a set day and a set amount. I feel like that will be confusing for opposing batters and good for Bundy. It does mean that Gallardo will have to pitch often.

I think Jimenez is the third best pitcher on the club and therefore should pitch a lot of innings. I also think free agents will be less likely to sign here unless those two get rewarded with wins.

Pip said...

The Astros tried the piggyback approach in Bo Porter's final season(2014, I think)
It didn't work at all, I don't recall why, and was abandoned before the all-star break.

Matt Perez said...

A number of teams have tried it but failed.

One worry was that there wasn't enough recovery time. Guys weren't able to go on three days rest. Another problem is that there wasn't any time to work on development and practice. Usually, the third day is when pitchers have a side session and on this schedule they wouldn't have it. You may need two long men in the bullpen to pull this one off.

steve said...

Too much change - this proposal has 4 parts. Deconstructing:
1)5 day rotation vs.5 game
2)relieve older starters (YG & UJ) early
3) start less used older starters more often
4)Long relief assigned rather than situational
Last two sound dodgey but the first 2 have real merit. The 5 man rotation evolved thru a non-data/analytic process. A 5 DAY rotation means using off days to skip the worst SP, the gain (maybe) trading off against stretch run fatigue. Worth the risk for this staff IMO.
Pulling old SPs early is marginally beneficial for YG; for UJ it's huge! UJ's 3 year OPS against is .690 for his 1st 25 pitches & then .680, .690, but .770 for pitches 76-100 (and a truly awful .930 on 76-100 the last 2 yrs.). He simply is an 80 pitch guy now. YG also falls off a bit: .720,.700,.740,.780 (baseline: all AL SPS were .720 on pitches 51-75 vs. .740 for 76-100.
So the moral is use off days for old guys and pull them out early. As an old guy, that sounds right.

Unknown said...
This comment has been removed by the author.
Unknown said...

If the Orioles did not make the move to go after Gallardo and Fowler and would have singed Chen for close to what they would have paid those two, the O's rotation would be pretty decent. Gauz and Tillman as the 1-2 punch at the top, with Chen as the 3rd. But instead they have Gallardo whom they also gave up a 1st round draft pick for. What a disaster of a move by the O's. The only way to recover from that may be to hope Bundy can stay healthy and possibly move into the rotation.

Pip said...

Signing Gallardo was indeed a disaster. All signs pointed to exactly what has happened: terrible production and injury.
I wish the Orioles had signed AJ Griffin, but somehow Dan missed that one.

Anonymous said...

Why speculate on something like this? Buck will never, ever do something this radical.

Matt Kremnitzer said...

I don't know how anyone could watch Bundy this year and feel confident with him working more than a couple innings at a time. Hopefully that changes, and anyway, it's at least a positive that he's still healthy.

Roger said...

Matt K, no one expects Bundy to do more now. We are all hopeful that, with some more work under his belt, he will be a rotation option later in the year. He has had so few innings in the last three years that he will need quite a bot of time to round into shape. He has shown enough for us to think he may be able to get better.

Also, to the others, the O's did pick up a draft pick in the Chen deal so that pretty much cancels the lost pick. But, yeah, Gallardo is pretty much bleah..... Matt Latos seems to be the big miss right now......

Matt Perez said...

Buck will likely never do anything that we suggest. But if another starter goes down, he might have to do something this radical. Or if starters deep below the 5 innings line.

I wanted Latos and Happ. Both have excellent ERAs and W-L records but their K% and BB% isn't looking good. I'm not sure either is going to end up as a bargain at this rate.

Anonymous said...

Here's my modified version of the "rotation of relay teams" idea.

* Use 12 pitchers and divide the pitching staff into 3 groups.
A: Tillman, Gausman, Gallardo, Jimenez, and Wilson
B: Wright, Worley, and Bundy
C: Brach, Matusz, O'Day, Britton

* Group A is your standard 5-man rotation. However, the assumption here is that the starter will be good for only 5 innings on average; the remaining innings would be covered by a combination of pitchers from Group B (who will be used as long relievers / "co-starters" depending on the situation) and Group C (short relievers).

- The best-case scenario would be getting 7+ innings from a starter. Should that happen, manage normally; the remaining innings would be covered via 1-2 pitchers from Group C.

- If the starter lasts for 6 innings, you can have someone from Group B cover the remaining 3 innings, or split them 2-1 with someone from Group C.

- If the starter manages to last only 5 innings as expected, no worries; someone from Group B will pick up the 2nd half of the game, with the expectation that he'll be unavailable for the next 4 games. In essence, he would be a "co-starter."

Here's an example of how the strategy might work over a span of 5 games.
Game 1: Tillman (5) + Wright (4)
Game 2: Gallardo (5) + Brach (2) + Matusz (1) + Britton (1)
Game 3: Jimenez (5) + Bundy (3) + O'Day (1)
Game 4: Gausman (5) + Worley (4)
Game 5: Wilson (5) + Brach (1) + Matusz (1) + O'Day (1) + Britton (1)

Notice that I've included a "worst-case scenario" in a non-blowout (using everyone from Group C in Game 5). Notice that no one from Groups A/B has to pitch on short rest, and none of the short relievers would need to pitch on consecutive days either unless absolutely necessary.

Obviously, the main drawback of this strategy is its inflexibility: For every single game, you'd have to pencil in not just the starter, but rather the names of the exact 2-4 pitchers who would be expected to be used in that game. Obviously, that means L/R match-ups are out of the question unless you're willing to burn the ultra-short 4-man bullpen. Also, you would practically need to give up early in case of blowouts: The starter from Group A would need to be his own mop-up pitcher in order to save the bullpen.

Matt Perez said...

I think Buck would be extremely unhappy to have Givens in Norfolk. He'd probably rather Wright in Norfolk and Matusz pitching multiple innings, presumably when he's given up on the Os chances of winning or when the Os have a safe lead of 8 runs or so.

To some extent, this already happens. I mean, the pitchers in Group C largely pitch the 7th-9th innings when the Os have a lead. The pitchers in Group B either pitch the sixth inning or later innings when the game isn't close. I feel like your plan would minimize the innings that Group C pitchers pitch while maximizing those of Group B pitchers.

Anonymous said...

I feel like your plan would minimize the innings that Group C pitchers pitch while maximizing those of Group B pitchers.

Fair point. The premise behind my idea was simply this: If the bullpen is being overused (in the sense that many short relievers have to pitch on 2-3 consecutive days on a regular basis), you should be able to avoid the situation by allowing longer outings. Perhaps a better middle ground might be something like this:

* Pencil in one A, one B, and two Cs for each game... but adjust the plan in-game.

- If it's a low-leverage situation, plan on using one A and one B so that the Group C pitchers can be saved for another game.

- For medium-leverage situations where the starter only lasts 5 innings, go with A + B + 2xC as originally planned.

- If it's a high-leverage situation, do the opposite: Feel free to use as many pitchers from Group C as you need in order to win a close game. (Just be prepared to handle the next game using the A+B strategy if necessary.)