Staying Ahead of the Curve: Why Conventionality is a Luxury of the Rich | The Process Report

Staying Ahead of the Curve: Why Conventionality is a Luxury of the Rich

Balling on a budget means that the Tampa Bay Rays constantly need to be seeking the next innovation in the game. As someone that follows the team closely it is always interesting to try to figure out the motivations for the various, inevitable moves. Sometimes it works out. Sometimes not so much. Most of the time the needle doesn’t move all that much from an individual decision, but there are enough tweaks and twerks that the total result can carry meaning. This is the team that brought back the ancient art of the overshift, and played the hell out of it while befuddled announcers cried, “WHAT IS THIS WITCHCRAFT!”

Extensive use of the platoon has been a hallmark since Stu Sternberg et. al. took the reins. The team has shown that it helps to have defensively flexible players to fully leverage this reliance . Some of those guys don’t take to their secondary position due to discomfort, inability, or outright refusal. Some don’t hit enough or become superfluous or just aren’t enough, but the team continues to find creative ways to build depth by targeting under-the-radar prospects. The last two years have seen high power, low contact, meh walk lineups that made more highlight reels, but still failed to actualize into a dynamic offense necessary to wage battle in the American League East.


(Table courtesy of

Decade-high Isolated Power (ISO) and Home Run totals over these past two years resulted in below average offenses, low run totals and massive team-wide strikeout rates. Compare this to the gloriest of glory years when from 2008 – 11 the team hit for solid power without making the whiff trade off, while also walking at exceptional rates. As a result, this offseason has seen the team’s annual sea change be a reversion to a more contact-oriented lineup throughout the system in an effort to develop the type of complementary or role players the team will need as brighter lights, perhaps, begin to shine. Having seen that the other path brings it’s fair share of travails it makes a lot of sense to walk away from the type of players that struck out aplenty even if they could get into one from time to time. While the 2008 squad had a guy like Carlos Pena that was about as three true outcome as you’ll find, they also had plenty of other skillsets that meshed together quite well. A team of Carlos Penas might be a lot of fun to be around, but it has proven to not be all that effective at generating runs.

The positional player side isn’t the only place where the team is seeking a return to the past. It wasn’t until the 1970’s where teams started to fully turn to the five-man rotation as a way to get through a season. The Rays are throwing on their fauxbacks and going old school by making an attempt to revert back to the four-man rotation that was en vogue prior to today’s norm. Neil Solondz recently made a comparison to Earl Weaver’s preferred way to break in young pitchers. Since I couldn’t agree more and think that he nailed this you should check out his background along with some of the more appealing aspects. The question becomes how a team can get through a season in this manner even if the Rays plan on running a modified version of the four-man where a call up backed by the bullpen is expected to function as the fifth starter, as needed? Especially, if the team looks to take their version of the quick hook up a notch?

Before getting too deep into planning this out we should probably have an idea of what a typical season looks like from a number of pitches thrown or batters faced perspective. To that end I have plotted the game by game figures for each of these perspectives with the # of pitches corresponding with the left-hand axis and the number of batters faced matching up with the right. The absolute first thing that jumps off the page is the outlier games that saw close to 280 pitchers or 75 batters faced. Since we’re only talking about a couple of data points that are so extremely out there it makes sense to realize right away that there is going to be a tripwire somewhere out there that is going to foul up this plan. The thing is, that sort of night is going to ruin any kind of plan you have and you have no idea when it will come. We should get it out of our head early that there is any sort of perfect plan that can cover all contingencies. Stuff is going to happen. Let’s take those two out and to be fair we’ll throw out the two lowest totals, too:

The rest of the data fits very neatly in a tighter band. I have added in what plus or minus one to two standard deviations looks like as all we really care about is the upper band. Most of the time we should expect to face fewer than 41 batters and it should take around 163 or fewer pitchers to seal the deal. Around 14% of the time you can expect to face as many as 45 batters requiring up to 181 pitches. This gives us a pretty solid target to shoot for.

Knowing around what we need for each game gives us a target, but we must acknowledge that the human body can only handle so much work over a short amount of time. To that end we need to establish some rules. Pitchers need to be thought of as filling three specific roles:

  • Two Times Through (2x) – These will be your usual starters, though they can also be available mid-game if on proper rest. The role is expected to face 18 batters or throw 80 pitches and should expect to be removed once they meet either of those conditions. Your very best starters can expect to face the first handful of batters a third time in appearances that haven’t been overly stressful or the pitcher seems to have great feel on a given day. Each 2x will always receive four days between appearances. You want to have at least five of these at all times, and a sixth would be ideal.
  • One Times Through (1x) – This role will mostly be expected to enter mid-game, usually to follow the starter, though the occasional nominal start will also be made or game closed out. This pitcher will face upwards of nine batters or 40 pitches, whichever comes first. Each of these pitchers will receive two full days off following an appearance. The team should have three of these at all times, though two would suffice when the team is fortunate enough to have a sixth 2x.
  • One Inning Pitched (1 IP) – These are your traditional setup men that can be trusted to go a full inning. Specialists can be carried when the team elects to forego a bench spot to carry a thirteenth pitcher, but generally the team will want a mix of lefties and righties that can avoid trouble when they have to work around an opposite-handed, good batter. These pitchers can go back to back, but never three days in a row, nor three appearances in four days.
  • 1x guys can become 2x guys and vice-versa if the team thinks the former is ready for increased responsibility or the latter could use a slight breather. 1 IP guys will nearly always remain in that role, but quick innings could occasionally see them go back out if and when the team needs it.


(Option information courtesy of

The summary above can be useful for getting an idea of which pitchers fall into which buckets. The pitchers highlighted in yellow would be your traditional starter types that would go around a max of 80 pitches, though I would be comfortable extending Chris Archer to a hard cap of 100 a lot of nights, while letting Blake Snell push to 90 when things are going well. Those in orange would be your 1x guys that will be vital to bridging the middle innings. Names in blue are your 1 IP guys. Note that given the choice of Chaz Roe and Daniel Hudson I elected to release the former. Each of the guys in red will be your reinforcements waiting at Durham. The team should be able to count on Yonny Chirinos, Anthony Banda, Jose Mujica and Hunter Wood to get through a lineup once upon being called into the rotation.

The team also has strong 1 IP candidates in Jose Alvarado, Diego Castillo, Andrew Kittredge, Ryne Stanek and Jaime Schultz. They face a tougher road to a call up simply due to the current 25-man roster candidates possessing no options collectively. In order for them to be called up the team will have to ride short on one of the other roles or have to play a man down on the bench. A disabled list stint for one of the penciled in names is also quite likely.

With parameters in place we can start to model what it might take for the team to get through the season utilizing this concept. I have broken this down by month below. You will want to open these far too small images in a new tab or window. On the left you will find the game date and opponent with a shot in the dark at which pitchers will be used to reach our goal of pitches thrown. On the right you’ll find the players currently on this hypothetical roster. Those in red are absolutely unavailable due to the rules put in place. Those in yellow are prospectively getting used, while those in white are available, though extra rest is always preferred. Don’t get too fixated on the names as it is more about the role being selected to get the team through the game. If an available guy gets used then he has to rest, but that means someone else is available down the line. This also goes for the callups as the team should look to stagger when these are made so that someone is always close to available considering their mandatory ten-day stint back in the minors.

Chris Archer gets the nod on Opening Day with each of the 2x guys to follow in order so that the very best pitchers get the chance to make the most appearances. There will be four separate games started by a 1x guy during this period and you can see the team maintains enough flexibility to follow that up with more 1x guys or a plethora of 1 IP. There would be a few callups necessary, but one of the advantages of this increased flexibility is that the team should be able to line guys up for promotion based on the upcoming opponent. This could be handedness splits, but also park-dependent as the team would have the option to run ground ball artists like Yonny Chirinos or Austin Pruitt when that presents as a better option.

In May the team would need to ensure they have the horses for a brutal ten-game road trip in the middle of the month that lends one offday before another long stretch of unbroken games. This stretch will be the first real challenge as you can see the number of unused, available pitchers shrinks considerably. The team will have to be very creative in cycling guys through to ensure health is maintained, but also that an effective run-suppression unit can be utilized. The ten-day disabled list might become rather enticing at this point as the most overworked pitchers are likely to be the 1 IP that may have to occasionally do an up-down when the team needs it and they’re under 20 pitches. A nice thing during this month is that no 1x guy needs to make a start.

June is shaping up to be the toughest month due to only three offdays, but also because the team is facing it’s most difficult stretch of opponents that they will see on the season. After having two months to sort out who isn’t going to work and which call ups are differentiating themselves the team will have more information to make better decisions. Players should also have started to acclimate into a routine as most are in a fairly normal rotation. Matt Andriese and occasionally one of the 1x pitchers are most likely to see the longest layoffs, but it also means they are available for when things get out of hand.

It’s a long grind into the All Star Break, but when the team gets to that Way Station of sorts they will be able to recharge the batteries on every single one of these options. This will be key for the second half as the team will have a strong idea of whether this concept is actualizing, and if it isn’t bringing success then the team will most likely be in a position to sell off assets of more worth to others than in-house. That could change things dramatically. Or maybe not at all. It’s no stretch to think the team could still be playing meaningful baseball, either, so with more information in hand the team will be able to choose how to proceed. That sort of flexibility may come in hand versus the rigidity of a traditional five-man rotation.

August features four offdays, and the team looks to have plenty of guys available as needed.

September means roster expansion so the team will be able to throw any number of pitchers in a single game. I have continued the concept here, but know that everyone is likely to see action, especially with important decisions to be made for 2019 whether the team is alive or not.

Plugging in the projections for number of pitches allotted in a game you can see that we’re well above where we were last year, which is a good thing for this experiment. With few occasions where the team was up against the wall to plug players in it is heartening to see that there should be even more availability as these estimates are well high over the course of the year. Part of that is that we’re using the max number of pitches allowed, but you hope to push a pitcher to that point most days.

Totaling everything up you can see how many appearances each pitcher will make with an average of 3.75 pitchers being used in each game. Pitches shows you how many each guy would throw under the maximum conditions, but this totals 26,330, while the team threw exactly 24,000 last year. Scaling the projections to last year gives you aPitches. You might notice that the 148.1 per game average matches up nearly identically with 2017 as a consequence. #math

Going to the right-hand box you can see projected ERA per the Fangraphs Depth Charts. As a team last year the starting pitchers allowed a .302 wOBA the first time they faced a batter, which went up a little over three percent to .312 the second time that batter came up. The much larger jump is when the pitcher faces a batter the third time, which yields an average of .345 wOBA allowed or around a 12.4% reduction in effectiveness compared to the average of the first two appearances. By forcing most of the starters to leave the game prior to that larger jump in ineffectiveness we should expect better individual results compared to the projections that do not account for this. This adjustment leads to the apERA column.

In the case of both Archer and Snell I discounted this adjustment by half thinking that in about half of their starts they will be used in the more traditional sense and the other half seeing a quicker hook. Maybe that isn’t aggressive enough, but you can see these two pitchers being a combined nearly 9 runs better over the course of the year when comparing the projected wRAA to the adjusted projected wRAA, which yields “d”. Those two alone should add roughly a win to the team, but the other pitchers will benefit from the full adjustment, and their improvement compiles, as well.

The 1x guys would see only the roughly three percent adjustment as they were unlikely to see an order three times, though now we’re taking it down to a maximum of one so those residuals are even smaller, but do add up. Altogether, I see the team projected to allow 637 runs based on my projected workload, which would have been third best last year trailing only the Los Angeles Dodgers and Cleveland Indians. Factor in the improved results due to less batter familiarity with the pitcher, and more importantly, less fatigue for that pitcher, and I see the team now allowing a mere 599 runs. This leads to around 3.70 runs allowed per game compared to the 3.93 as projected by Depth Charts. It doesn’t sound like much shaving 0.23 runs per game off, but that is a run every four games and roughly 38 on the year. Adding those roughly four wins in free agency would cost something like $34,000,000. You can see why this is worth the effort.

You can call it a four-man rotation, with or without the preceding “modified,” but in reality this is the team finding a way to use the 15 to 20 pitchers that occupy the 40-man roster most effectively. The rotation will look fairly normal other than the occasional spot start or pitcher getting pulled earlier than your grandma would like. The team will rely on a revolving door of useful swingmen in the middle innings knowing that they aren’t always going to have their best stuff, but that they will have to eat it anyway to keep the ship afloat. This will likely be difficult for many players to process, but the team has an opportunity to show the league the next avenue to squeezing blood from a stone. Those that are here to win will get it. This latest experiment should be extremely interesting to follow over the course of the year, and if it falters, well the team isn’t all that much worse off than if they had stayed conventional. If it works, you can rest assured that the rest of the league will follow suit quickly. As they did with shifts, platoons, super utility guys and everything else the Rays got to first, if not for long.