Home | Webstore
Latest News: OOTP 25 Available - FHM 10 Available - OOTP Go! Available

Out of the Park Baseball 25 Buy Now!

  

Go Back   OOTP Developments Forums > Prior Versions of Our Games > Out of the Park Baseball 20 > OOTP 20 - Historical Simulations

OOTP 20 - Historical Simulations Discuss historical simulations and their results in this forum.

Reply
 
Thread Tools
Old 04-19-2019, 12:28 AM   #1
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
Pitcher usage in the pre-reliever era, part II

THE SET-UP

This series of posts will continue my examination (see this thread) of the way that OOTP handles pitchers in the pre-reliever era (roughly pre-1940). Briefly summarized, I believe that the most accurate results possible under the game's current pitching framework can be achieved by making the number of pitchers on a staff equal the number of pitchers in the pitching rotation. That would mean that all (or at least six) members of the staff will be starters. Prior to about 1915 there shouldn't be any relievers, and between approximately 1915 and 1940 there should only be a few relievers in certain limited circumstances.

In this test, I played as the Detroit Tigers in the 1922 season. The reason I chose the Tigers is that they had a relatively straightforward pitching staff that year. The team used only eleven pitchers during the entire season, and of them three pitched a total of 30 innings, meaning that the top eight pitchers accounted for 98% of all innings pitched for the club. Since I planned to limit pitching staffs to eight members, I felt that would make it easier to compare the game's results with the stats for the real-life team.

As mentioned, I changed the default settings, which are set at nine-man pitching staffs and 25-man rosters. Instead, I set pitching staffs at eight men and rosters at 23. I believe this is more realistic, even though teams in 1922 were allowed 25 men on their rosters. In practice, I don't think any teams carried this many players, and the fact that the Tigers in 1922 got along for most of the season with eight pitchers seems to bear this out (in fact, they probably carried fewer than 23 - only 16 non-pitchers suited up for Detroit that season).

I set the league's rotation size at six and the rotation mode at "start highest-rested." I ranked the starting rotation in order of the number of starts that each pitcher recorded in real life. So Herman Pillette headed the rotation, even though OOTP wanted to make him into a reliever (must have something to do with his ratings). Here's the starting rotation, with the number of real-life game starts for each pitcher:

Herman Pillette - 37
Howard Ehmke - 29
Red Oldham - 28
Hooks Dauss - 25
Ole Olsen - 15
Syl Johnson - 8

Having eight-man staffs meant that two pitchers would necessarily become relievers. In this case those were Bert Cole and Lil Stoner. It might be argued that Johnson was really a reliever - he had 19 relief appearances in 1922 as opposed to only eight starts, but Cole and Stoner were likewise predominantly used in relief (18 relief appearances and 5 starts for Cole, 10 and 7 for Stoner). Plus, I was tickled that I had Olsen & Johnson at the bottom end of my rotation. So sue me.

Ty Cobb, as the skipper of the 1922 Tigers, was slightly ahead of his managerial counterparts in his use of relievers. His club ranked last in the AL in complete games and trailed only the Browns in the number of "retrospective saves." That could be the result of some innovative thinking on the Georgia Peach's part, but I believe it had more to do with the generally mediocre quality of his staff. In that respect, he probably got more out of his pitchers than he is generally given credit for, as his team finished with a 79-75 record, good enough for third place. The fact that Detroit used relievers more often than normal, I thought, might actually yield more realistic results given the limitations of the OOTP pitching model.

There is no pre-set strategy profile for Cobb, and there are only a smattering of managerial profiles suitable for 1922. By default, I set any team that did not have a manager with a profile (such as Connie Mack and John McGraw) to Pat Moran, who was the manager for the Reds. Moran was also something of a leader when it came to the use of relievers, but I'm not sure if his strategy profile really reflects that fact. According to the manager background file: "Pat wants his starters to pile up innings but will remove his relievers quickly if necessary." Whatever that means.

CONTINUED
joefromchicago is offline   Reply With Quote
Old 04-19-2019, 12:29 AM   #2
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
THE SIMULATION (SIM1)

I followed a "set it and forget it" style of managing for the most part. That was made easier by the fact that my pitchers suffered no injuries of any kind during the season, which means that I only used eight pitchers in total. Their stats (games-games started-innings pitched) in the sim and in real life are as follows:

Hermann Pilette (sim): 38 G, 32 GS, 256.2 IP
Hermann Pilette (RL): 40 G, 37 GS, 274.2 IP

Howard Ehmke (sim): 35 G, 31 GS, 254 IP
Howard Ehmke (RL): 45 G, 29 GS, 279.2 IP

Red Oldham (sim): 34 G, 29 GS, 226 IP
Red Oldham (RL): 43 G, 28 GS, 212 IP

Hooks Dauss (sim): 38 G, 22 GS, 206 IP
Hooks Dauss (RL): 39 G, 25 GS, 218.2 IP

Ole Olsen (sim): 27 G, 26 GS, 184 IP
Ole Olsen (RL): 37 G, 15 GS, 137 IP

Syl Johnson (sim): 15 G, 14 GS, 87 IP
Syl Johnson (RL): 29 G, 8 GS, 97 IP

Bert Cole (sim): 42 G, 0 GS, 74.1 IP
Bert Cole (RL): 23 G, 5 GS, 79.1 IP

Lil Stoner (sim): 46 G, 0 GS, 83.1 IP
Lil Stoner (RL): 17 G, 7 GS, 62.2 IP

The biggest differences, of course, are the stats for the two pitchers who were used exclusively in relief in the sim: Cole and Stoner. Although their IP numbers are surprisingly close to their real-life totals, the way that they racked up those innings was definitely not. In 1922, only four pitchers in the AL finished with more than 45 appearances, and they all had at least 32 starts. Hub Pruett of the Browns led the AL with 31 relief appearances, but even he started eight games.

Had any of my starting six suffered an injury, Cole and Stoner would surely have had a chance to get in a few starts. But that's not how Cobb used his pitchers in 1922. He also had a fairly stable pitching staff - Dauss may have started the season injured (he didn't make his first appearance until April 24, twelve days after the opener) - but in general Cobb used his top eight pitchers fairly consistently throughout the season.

What Cobb didn't do was use a six-man rotation, or a rotation of any sort. By the second half of the season he would occasionally fall back on a starting sequence of Pillette-Ehmke-Oldham-Dauss, but that was by no means a regular occurrence. In general, I think Cobb did what most managers at that time did: he picked the best pitcher who appeared rested and ready that day to start. There may also have been some lefty-righty percentages at play. Oldham, the only left-hander who started consistently (Cole was the other lefty on the staff) had ten of his 28 starts against Washington, a team whose best hitters were all left-handed.

With only two men in the bullpen, the starters in the sim all had chances to relieve. Dauss's sim numbers are surprisingly close to his real-life totals: I'm not sure what was going on there, but it's a least a promising sign. Oldham appeared more often as a reliever in real life than in the sim, probably because he was reliable and left-handed. Johnson, at the bottom of the rotation, still got 14 starts, six more than in real life. And despite having two pitchers who were strictly limited to relief, the sim Tigers still compiled more complete games (72) than the real-life Tigers (67). That's in line with my previous simulations, which also saw this paradoxical result.

So OOTP still can't replicate the way pitchers were used in the pre-reliever era - at least not with my settings. Are the out-of-the-box settings any better? We'll see....

CONTINUED
joefromchicago is offline   Reply With Quote
Old 04-19-2019, 12:31 AM   #3
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
THE SIMULATION (SIM2)

As mentioned before, the OOTP pitcher settings for 1922 are five-man rotations and nine-man staffs. Maybe these could actually yield more realistic results. To test that hypothesis, I again ran the Tigers. I still ranked the starters according to the number of starts they had historically, which gave me the following staff:

Herman Pillette (SP)
Howard Ehmke (SP)
Red Oldham (SP)
Hooks Dauss (SP)
Ole Olsen (SP)
Syl Johnson (RP)
Bert Cole (RP)
Ken Holloway (RP)
Carl Holling (RP)

In this sim, Lil Stoner was the odd man out. He didn't make the opening-day roster, for reasons that I can't remember (probably because he stinks), although he did get called up to the bigs twice because of injuries to Holloway and Holling. In real life, Holling and Holloway combined for 10.1 innings pitched. Once again, I used "set it and forget it" for the starting rotation, and once again none of my starters suffered any serious injuries. As a result, the stats for this sim (sim2), and their comparison with the previous sim (sim1) and real life, were as follows:

Hermann Pilette (sim2): 40 G, 38 GS, 316.1 IP
Hermann Pilette (sim1): 38 G, 32 GS, 256.2 IP
Hermann Pilette (RL): 40 G, 37 GS, 274.2 IP

Howard Ehmke (sim2): 41 G, 33 GS, 292 IP
Howard Ehmke (sim1): 35 G, 31 GS, 254 IP
Howard Ehmke (RL): 45 G, 29 GS, 279.2 IP

Red Oldham (sim2): 39 G, 31 GS, 253 IP
Red Oldham (sim1): 34 G, 29 GS, 226 IP
Red Oldham (RL): 43 G, 28 GS, 212 IP

Hooks Dauss (sim2): 33 G, 29 GS, 242 IP
Hooks Dauss (sim1): 38 G, 22 GS, 206 IP
Hooks Dauss (RL): 39 G, 25 GS, 218.2 IP

Ole Olsen (sim2): 27 G, 23 GS, 157 IP
Ole Olsen (sim1): 27 G, 26 GS, 184 IP
Ole Olsen (RL): 37 G, 15 GS, 137 IP

Syl Johnson (sim2): 5 G, 0 GS, 7 IP
Syl Johnson (sim1): 15 G, 14 GS, 87 IP
Syl Johnson (RL): 29 G, 8 GS, 97 IP

Bert Cole (sim2): 50 G, 0 GS, 77.1 IP
Bert Cole (sim1): 42 G, 0 GS, 74.1 IP
Bert Cole (RL): 23 G, 5 GS, 79.1 IP

Lil Stoner (sim2): 11 G, 0 GS, 25.2 IP
Lil Stoner (sim1): 46 G, 0 GS, 83.1 IP
Lil Stoner (RL): 17 G, 7 GS, 62.2 IP

Carl Holling (sim2): 12 G, 0 GS, 15.2 IP
Carl Holling (sim1): did not play
Carl Holling (RL): 5 G, 1 GS, 9.1 IP

Ken Holloway (sim2): 6 G, 0 GS, 9.2 IP
Ken Holloway (sim1): did not play
Ken Holloway (RL): 1 G, 0 GS, 1 IP

Were the results for sim2 any more realistic than those from sim1? Well, it could be said that the top three pitchers compiled too many innings in sim2, but then it could also be said that they compiled too few innings in sim1. I was actually surprised by the number of relief appearances made by the top starters in sim2. With four relievers in the bullpen, I would have thought that the AI would use the starters in relief only sparingly. That, however, wasn't the case. Oldham and Olsen, for instance, were used more often in relief in sim2 than in sim1.

The starts, of course, were not as widely distributed in sim2 as they were in sim1. The biggest loser in this regard was Johnson, who went from 14 GS and 83.1 IP in sim1 to zero GS and 7 IP in sim2. That sort of result should be expected. In the game, Johnson isn't really good enough to be either a regular starter or a regular reliever. With a six-man rotation, he gets some IP just because he's the sixth man in the rotation. With a five-man rotation, however, he doesn't really have a role to play. His innings, then, largely went to the top four men in the rotation, who collectively accumulated 17 more starts in sim2 than they did in sim1.

The five starters also accumulated more complete games in sim2 (82) than the six starters did in sim1 (72) and more than the 1922 Tigers did in real life (67) - and this despite having four relievers in the bullpen. That may have had something to do with the strategy pre-sets, which I didn't change from the out-of-the-box settings (neutral for everything). Would there have been fewer CGs in sim2 if the manager was Pat Moran, as was the case in sim1? I don't know, but it seems odd that the AI would actually use relievers more often by employing the strategy of a 1920s manager. Something's off here, and I'm not sure what it is.

One other stat stands out. Bert Cole dominated the bullpen in sim2. In sim1, Cole shared bullpen duties with Stoner, but in sim2 Cole had more IP than the other four relievers put together. Cole finished 32 games and saved 8. Not only were these numbers unrealistic, but they indicate that the AI was using Cole either as a stopper or a closer, roles that didn't really exist in 1922. So that's unrealistic on a couple of levels. Indeed, despite the fact that the game setting for "use of closers" is "very rarely," some relievers on other teams (though not on Detroit) were designated as closers in sim2. That just shouldn't be happening.

CONTINUED
joefromchicago is offline   Reply With Quote
Old 04-19-2019, 12:33 AM   #4
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
OBSERVATIONS

So which is the more accurate way to play the game? In both sims, the numbers for the top four starters were, I think, within an acceptable range relative to their real-life numbers. The biggest deviations, then, come with the fifth and sixth starters, and also with all of the relief pitchers. In an era when every pitcher was considered a starter, it stands to reason that all of the pitchers on a staff should get a chance to start. Expanding the rotation to six, then, at least spreads out starts in a more realistic fashion.

That still doesn't remedy the problem with the bullpen. As I said in my previous thread, OOTP took a big step forward when it allowed starters to relieve. Now it needs to take the next step and allow relievers to start. Theoretically, they can - a reliever can have "emergency SP" set as his primary role. The problem, though, is that the AI never seems to do that. It won't solve the problem if only one team in a league (the one that's being run by the human at the keyboard) ever employs relievers as emergency starters. Furthermore, "emergency SP" doesn't accurately describe how some of these "bottom of the staff" pitchers were actually used, unless "second game of a double-header" constitutes an "emergency."

The other big problem with the bullpen is that the AI can't seem to shake the idea that it's 2019 and it has to use relievers in the same way that they're used today. Very few managers in the pre-reliever era thought of their bullpens in that way. If a pitcher was used mostly in relief, it was because he wasn't good enough to start. The bullpen, then, was mostly a purgatory for failed starters. As a result, a manager would look to his relievers less with a sense of optimism than with a foreboding of doom. That's why most managers, when a game was on the line in the late innings, would call on one of their established starters to come into the game. Perhaps the most famous relief appearance in the 1920s shows this way of thinking. In the 1926 World Series, Rogers Hornsby, manager of the Cardinals, called on Pete Alexander to come into the seventh inning of game seven, even though Alexander had pitched a complete game the day before and the Cardinals had four pitchers rested and ready in the bullpen.

True, some managers were already experimenting with relievers as early as the 1910s. In 1922, Branch Rickey, manager of the Cardinals, had two pitchers - Lou North and Clyde Barfoot - who could legitimately be called relief specialists, although they still weren't used like modern-day closers. In 1925, Fred Marberry became the first pitcher to record 50 or more appearances in a season without a start. Presumably, that's the sort of thing that should be replicated through strategy settings. Some managers, like John McGraw and Bucky Harris and Branch Rickey, were more likely than their peers to use relievers. For the rest of the managerial fraternity, though, the bullpen was a place of exile.

So one thing that I think needs a closer look is managerial strategies. This is something that I didn't address in my previous thread, but the more I think about it, the more important this issue becomes. Right now, I think a lot of the pitcher-handling strategy is dictated by the "General Strategic Tendencies" listed in the League Settings/Stats & AI page, specifically the "hook" settings, the "use of relievers/closers" settings, and the "pitcher stamina" setting. These settings are supposed to work together to create a historically realistic pitching environment for the season. I'm not convinced that they do.

As we can see from these 1922 simulations, even a setting of "very rarely" for use of relievers/closers doesn't prevent pitchers from relieving far more often than they did in real life. The "hook" settings, on the other hand, might have the opposite effect of keeping starters in too long. Thus the paradoxical result of having too many relief appearances and too many complete games.

Of course, one of the major culprits behind the over-use of relievers is the fact that they're in the game in the first place. That's the unavoidable result of having a pitching model built around the modern notion of the rotation. As long as rotations are limited to six pitchers, any staff of seven or more pitchers will have to include at least one reliever. There's no way around that, as far as I can see. As long as the game forces some pitchers to become relievers in the pre-reliever era, I don't think any amount of fiddling with the general strategic tendencies will be able to compensate for that.

CONTINUED
joefromchicago is offline   Reply With Quote
Old 04-19-2019, 12:37 AM   #5
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
CONCLUDING REMARKS

Nothing in these simulations has convinced me to reject or revise the recommendations I laid out here. The problems that were evident in OOTP19, in other words, are still present in OOTP20 (not that I expected any changes - there was no indication from the developers that they were working on, or indeed interested in, this issue). I still don't know how the hook ratings play a role in this, but it's clear to me that, if the hook ratings were intended to compensate for the over-use of relievers in the pre-reliever era, they have not succeeded.

As I noted above, one area that I did not cover before was managerial strategy. In sim1, I picked strategies from the handful of historical managers that are included with the game, as I wanted to see if that, together with my recommended settings for pitching staffs, would yield more historically accurate totals. The results were, let's just say, mixed.

In sim2, I didn't change any of the manager settings, because I wanted to see what the game would do on its own. From what I can gather, if a manager's strategy isn't set by the user, then the AI will set it. That means that it will either choose a strategy from the list of nine generic strategy presets available in the game or else it will choose "conventional," which is not an option that can be chosen from among the strategy presets. I'm not sure what a "conventional" strategy is, or if it has some connection to the league's general strategic tendencies. It appears that different managers who use "conventional" strategies don't use identical strategies, but all of the slider settings are within a tick or two of neutral, which suggests that there's some slight randomization of the "conventional" strategy settings (that, by the way, also appears to be the case with the generic settings, e.g. tactician, smallball, etc.). In short, I don't know if a "conventional" strategy in 1922 is the same as a "conventional" strategy in 2019.

In general, there needs to be more attention paid to managerial strategies, especially in the transitional period (approx. 1915-1940), when managers were experimenting with the use of pitchers who were primarily relievers. Some of the statistical outliers, such as Fred Marberry, could be handled solely through managerial strategy settings. In other words, having Bucky Harris as Marberry's manager should yield a different result than if Connie Mack or John McGraw was his manager. Indeed, Harris has much higher "hook" ratings than his peers, so this is at least one way to make the game use some of these "proto-relievers" in a more historical fashion (Harris, though, wasn't in my sims - he started his managerial career in 1924).

Managerial strategies, however, are something of a black box. Apparently, the person who created the game's strategies is no longer active in the OOTP community, and hasn't been since 2015. How he arrived at his numbers or what formula he might have used, then, seems to be a mystery that may be unsolvable. That's a shame, because the managerial strategies need some work. OOTP simplified the strategy sliders a few versions ago, reducing the number of game situations from 36 to 16. In one sense, that was a welcome change considering the amount of time that could be spent adjusting all of those sliders, but a lot of strategic nuance was, I think, lost in the process. For instance, instead of separate strategies for when a team is behind by a run or ahead by a run, now the only option is to set a strategy for when the team is either behind or ahead by a run. That might not seem like a big deal, but in the late innings it matters a lot whether your team is looking at a one-run lead or a one-run deficit. I'm not sure if those more in-depth strategic choices can be reintroduced, or if they're gone for good.

In addition, while the historical managers will have different strategic tendencies based on what inning it is, it appears that all of the generic strategies are the same for all innings and all situations. So generic AI managers will bunt or hit-and-run or use relievers the same way in the first inning as they do in the ninth inning, regardless of the score. The only way around this, apparently, is to do what I did in sim1: assign a historical manager's strategy profile to each team. It seems to me that the user shouldn't have to do that in order to achieve more realistic results from the game.

But any changes to the strategy tendencies, while welcome, would still just be tinkering at the margins of the problem. As long as the game insists upon having relievers in the pre-reliever era, it doesn't really matter very much what the AI managers' overall strategies are. If they have relievers, they'll use them like relievers. So, as long as the game's pitching model is based on pitching rotations, the first step is to make the rotation equal the size of the staff. These 1922 simulations reaffirm my belief that the game, as it is currently configured, can't handle pitching staffs in the pre-reliever era in a historically realistic fashion.
joefromchicago is offline   Reply With Quote
Old 04-19-2019, 01:00 AM   #6
Garlon
Hall Of Famer
 
Join Date: Jun 2004
Posts: 3,736
This is interesting. I go with a 4-man strict rotation from 1920-1935 and 3 relievers and a roster of 21 players. I am noticing that even though I set limits to 3 relievers, if I allow there to be say 16 position players and 7 relievers in the strategy settings that the computer will sometimes ignore this and put 4 or 5 relievers on the roster. I think the computer will not attempt to use fewer than 14 position players though. So you might want to try a roster of 21 players with 14 position players, 6 SP with start highest rested as the mode and 1 RP in the settings and allow starters in relief and see if you get better results like that.

You may want to try adjusting the Pitcher Stamina as well.

Last edited by Garlon; 04-19-2019 at 01:02 AM.
Garlon is offline   Reply With Quote
Old 04-19-2019, 09:52 AM   #7
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
Quote:
Originally Posted by Garlon View Post
This is interesting. I go with a 4-man strict rotation from 1920-1935 and 3 relievers and a roster of 21 players. I am noticing that even though I set limits to 3 relievers, if I allow there to be say 16 position players and 7 relievers in the strategy settings that the computer will sometimes ignore this and put 4 or 5 relievers on the roster. I think the computer will not attempt to use fewer than 14 position players though.
I wouldn't use the strict rotation setting before the 1950s. No team employed pitching rotations before then, and even then it was regarded as something of a curiosity. "Start highest rested" is, I think, a more historically accurate strategy for this era, although I plan on going back through the game records to see exactly what OOTP considers "rested." In 1922, for instance, there were a couple of instances where a Detroit pitcher started games on consecutive days. I don't think that can happen in OOTP, even if the pitcher has high stamina and pitches only one or two innings in the first game.

I would think that having four-man rotations would result in far too few starters pitching far too many innings. In 1922, only four pitchers started 38 games or more (roughly 25% of their team's games) and only three pitched more than 300 innings. I'm happier using six-man rotations and "start highest rested."

I haven't noticed the AI adding relievers beyond the bullpen limit set in the League Settings/Stats & AI page, but I wasn't really paying much attention to the AI teams. I'll have to go back and look at that. As a general rule, the AI used more pitchers than I did, but I think that is, in part, due to the AI aggressively recruiting free agents and promoting them when rosters expanded on September 1.

Quote:
Originally Posted by Garlon View Post
So you might want to try a roster of 21 players with 14 position players, 6 SP with start highest rested as the mode and 1 RP in the settings and allow starters in relief and see if you get better results like that.
I think a 23-man roster is more historically accurate for 1922, although I do think Detroit and probably some other teams used fewer players on a regular basis. I used 21-man rosters for my 1916 replay, although I used eight-man pitching staffs instead of seven.

I haven't seen much evidence on the number of players carried by teams in this era. There were roster limits in place and we know what those were, but teams routinely suited up fewer players than they were allowed (a roster minimum wasn't established until 1977). Just looking at the stats, it's pretty clear to me that the Tigers used eight pitchers for the vast majority of the 1922 season. Some teams may have used seven. I doubt any team used fewer than seven or more than eight, at least for any extended period of time. Of course, pitching staffs grew over time. In the early years of the twentieth century, six-man staffs were the norm.

Quote:
Originally Posted by Garlon View Post
You may want to try adjusting the Pitcher Stamina as well.
That's certainly one possibility. The problem that I encountered is that the results suggest that pitcher stamina is too high and too low - too high because too many pitchers are completing games, and too low because the AI is using too many relievers. There are a lot of moving parts to take into account here, and changing one might have an unintended effect somewhere else.
joefromchicago is offline   Reply With Quote
Old 04-19-2019, 01:10 PM   #8
Spritze
OOTP Historical Czar
 
Spritze's Avatar
 
Join Date: Dec 2001
Location: Bothell Wa
Posts: 7,254
ya gotta throw in changes to pitch counts per inning over the years as well?
__________________
It's madness, madness, I tell you! For the love of God, don't do it!
Spritze is offline   Reply With Quote
Old 04-19-2019, 03:21 PM   #9
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
Quote:
Originally Posted by Spritze View Post
ya gotta throw in changes to pitch counts per inning over the years as well?
Ah, Spritze, I am always amused by your delightfully dry wit, but sometimes I can't be sure if you're kidding or not. In case you're serious, I'll state that I didn't pay any attention to pitch counts in my sims. As far as I know, that information is not available for 1922.
joefromchicago is offline   Reply With Quote
Old 04-22-2019, 02:24 AM   #10
Spritze
OOTP Historical Czar
 
Spritze's Avatar
 
Join Date: Dec 2001
Location: Bothell Wa
Posts: 7,254
for 1922 you will notice that IP per week is hugely different which is also a form of pitch counts.

if you make your 2 bullpen guys openers and set it to very often they will get about 5 - 6 innings per start if they got the stamina
__________________
It's madness, madness, I tell you! For the love of God, don't do it!
Spritze is offline   Reply With Quote
Old 04-22-2019, 09:24 AM   #11
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
Quote:
Originally Posted by Spritze View Post
for 1922 you will notice that IP per week is hugely different which is also a form of pitch counts.
I suppose that's true, but unless the AI uses pitch counts to spread out starts or the user has the ability to set that in the league settings, there's really no point in talking about it.

Quote:
Originally Posted by Spritze View Post
if you make your 2 bullpen guys openers and set it to very often they will get about 5 - 6 innings per start if they got the stamina
That's a very interesting option. I might try that just to see what happens. Most "relievers" in the pre-reliever era were actually starters who, for one reason or another, were relegated to the bullpen. They should, therefore, have better stamina than the typical modern-day reliever. Is there any chance for an opener to pitch a complete game, or are they always pulled early?
joefromchicago is offline   Reply With Quote
Old 04-22-2019, 10:00 AM   #12
scott1964
Hall Of Famer
 
Join Date: Oct 2005
Location: Denver
Posts: 5,551
Blog Entries: 19
In 1923 with my Washington Senators I have a pitching staff of 8 with 7 starters and a single reliever. Rotation is 4 man
__________________
This just feels more like waiting in line at the Department of Motor Vehicles.

PETA.....People Eating Tasty Animals.

scott1964 is offline   Reply With Quote
Old 04-22-2019, 10:58 AM   #13
Spritze
OOTP Historical Czar
 
Spritze's Avatar
 
Join Date: Dec 2001
Location: Bothell Wa
Posts: 7,254
Quote:
Originally Posted by joefromchicago View Post

That's a very interesting option. I might try that just to see what happens. Most "relievers" in the pre-reliever era were actually starters who, for one reason or another, were relegated to the bullpen. They should, therefore, have better stamina than the typical modern-day reliever. Is there any chance for an opener to pitch a complete game, or are they always pulled early?
the pitchers are not good enough to pitch complete games as a general rule. it would take a lot of tryin and cryin to see what happens
it might be easier to just have unlimited starters or use 7 day lineups or even real lineups?
__________________
It's madness, madness, I tell you! For the love of God, don't do it!
Spritze is offline   Reply With Quote
Old 04-22-2019, 11:55 AM   #14
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
Quote:
Originally Posted by scott1964 View Post
In 1923 with my Washington Senators I have a pitching staff of 8 with 7 starters and a single reliever. Rotation is 4 man
What sort of results have you been getting?
joefromchicago is offline   Reply With Quote
Old 04-22-2019, 12:01 PM   #15
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
Quote:
Originally Posted by Spritze View Post
it might be easier to just have unlimited starters or use 7 day lineups or even real lineups?
Unlimited starters? Is that an option?

I could certainly use historical line-ups if my intention was to re-run historical seasons. But my main goal is to be able to run fictional leagues that reflect historical stats. That's my main interest. But to get fictional leagues to work properly, we first have to get historical leagues to work properly.
joefromchicago is offline   Reply With Quote
Old 10-15-2019, 01:57 AM   #16
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
My investigations continue.

For this simulation, I chose the 1920 White Sox. The Sox had a very stable pitching staff that year, which is ideal for testing the OOTP pitching model. The Sox staff was also unusual in that it used five starters almost exclusively - Red Faber, Eddie Cicotte, Lefty Williams, Dickey Kerr, and Roy Wilkinson. Only two other pitchers - Shovel Hodge and Joe Kiefer - started for the Sox that year, and they started a total of three games and pitched a total of 24.1 innings. No other team that year used so few starters on a consistent basis, although two clubs - the Giants and the Pirates - came close. The Giants had only five games not started by their top five starters, while the Pirates had eight.

As with the previous sim, I set the league rotation size at six and the rotation mode at "start highest-rested." I set the number of relievers to one, but I also set the number of position players to 15 and the roster limit at 23. With seven men on the pitching staff and 15 position players, that meant that there was one "free" spot on each roster that the AI could fill any way it wanted. It turned out that the AI desperately wanted to fill that spot with another reliever.

The AI initially set up the White Sox rotation with only five starters - Faber, Cicotte, Williams, Kerr, and Wilkinson. That made sense, as the Sox really didn't have anybody else who could start, and it was historically accurate. As noted above, the Giants and Pirates could also have been set up with five-man rotations, but the AI stuck with the strategy pre-sets and gave them six-man rotations. That wasn't surprising, although it was a bit disappointing. What was surprising is that some AI teams routinely carried three or even four relievers during the season. I'm not sure why the AI rigidly adhered to the rotation pre-set but ignored the bullpen-size pre-set. Maybe those teams couldn't fill out their quota of 15 position players, but that's just a guess.

I wanted the rotation to match how the pitchers ranked in terms of their innings pitched, which would have meant a rotation of Faber-Cicotte-Williams-Kerr-Wilkinson. I forgot to make that change, so the initial rotation (set by the AI) had Cicotte first and Faber second. I switched them after the opening game, but it took until about mid-June for the rotation to settle into the order that I wanted. Here are the stats for games started, with real life (RL) listed first and sim second.

Faber RL-39 sim-34
Cicotte RL-35 sim-34
Williams RL-38 sim-33
Kerr RL-27 sim-29
Wilkinson RL-12 sim-23

What we see here is what we saw in my previous simulations: the pitchers at the top of the rotation start too few games, while those at the bottom start too many. But the effect is exaggerated here because there are only five pitchers in the rotation. In real life, the top three starters for the Sox started 73% of the team's games, which I would guess is the top mark in the majors in 1920 but nevertheless not unusual - the Cardinals' top three came in at 72% and Cleveland's top trio clocked in at 71%. Teams would typically rely on three or four pitchers for the vast majority of starts, while the rest of the staff would pick up starts here and there on an as-needed basis. OOTP, on the other hand, just can't recreate that method of doling out starting assignments. It doesn't do a very good job with a six-man rotation, but it seems to do even worse with a five-man rotation. That calls for a closer look...
joefromchicago is offline   Reply With Quote
Old 10-15-2019, 02:12 AM   #17
joefromchicago
Hall Of Famer
 
joefromchicago's Avatar
 
Join Date: Jun 2011
Posts: 3,630
What is it about a five-man rotation in the pre-reliever era that totally breaks the OOTP model? First, let's take a look at the way that the game handled starts. As mentioned before, I stuck with a five-man rotation throughout the season, and none of my pitchers missed any time due to injuries. The chart below shows which pitchers started which games during the season:

1 - Faber
2 - Cicotte
3 - Williams
4 - Kerr
5 - Wilkinson
0 - off-day
(xx) - double-header

APRIL
2 1 3 0 4 2 1 3 5 4
2 1 3 5 4 2 0

MAY
1 3 4 2 5 1 0 (34) 2 0
1 5 3 4 1 0 2 3 4 1
5 2 3 4 1 5 2 3 4 1
(52)

JUNE
3 4 0 1 (25) 3 0 1 2 4
3 5 1 2 3 4 5 1 2 3
4 5
1 0 2 3 4 1 2 3

JULY
5 4 1 2 (35) 4 0 1 2 3
0 4 1 2 3 5 4 1 2 3
5 4 1 2 3 5 0 1 2 3
4


AUGUST
5 1 2 3 4 5 1 2 3 4
5
1 2 3 4 0 0 1 2 3
4 5
0 1 0 2 3 4 0 1
2

SEPTEMBER
3 0 4 1 2 (35) 4 0 1 2
3 4 5
1 0 2 3 4 1 5
2 0 3 1 0 2 0 0 0 0

OCTOBER
1 2 0 3

One thing that we can see from this chart is how often the 12345 pattern is repeated. Once the rotation got settled in mid-June (as explained in greater detail above), the 12345 pattern occurred seven times (in bold above). How many times did that 12345 pattern occur in real life? Never. A big part of the reason for that is that Wilkinson didn't start at all after July 31. It looks like the Sox manager, Kid Gleason, began the season with Wilkinson pencilled in as the fifth starter, but as the pennant race heated up he relied exclusively on his top four. That may have been Gleason's style as a manager, or he could have resorted to it out of necessity. Wilkinson wasn't very good, and the Sox always seemed to be pitcher-poor during that era (they had the same problem in 1919, only worse).

I very much doubt that an AI manager could make the same decision as Gleason and go to a four-man rotation halfway through the season. But even a human manager would find it difficult to make that switch. That's because a four-man starting staff would gradually disintegrate due to exhaustion. Even the Sox top four, who all have stamina ratings of 80 on a 20-80 scale, wouldn't be able to tolerate that much extra work. In OOTP, with pitcher stamina set at "high" (the default setting), pitchers still usually get four days of rest between starts. Here's a chart that shows how much rest each of the Sox starters got between starts (DR=days rest):

Code:
FABER         3DR-2 4DR-26 5DR-4 6+DR-1
CICOTTE       3DR-3 4DR-24 5DR-4 6+DR-2
WILLIAMS      3DR-3 4DR-21 5DR-6 6+DR-2
KERR          3DR-3 4DR-15 5DR-7 6+DR-3
WILKINSON     3DR-1 4DR-10 5DR-1 6+DR-10
So the overwhelming majority of time, a starter had four days of rest between starts. Needless to say, that's not historically accurate. Faber, for instance, started most of his games on two or three days of rest. Only eight of his 39 starts were preceded by four or more days of rest. Cicotte and Williams were the same. In OOTP, however, they can take it easy - four days of rest is the norm, not the exception.

But note that while a starter typically gets four days of rest, he doesn't need four days of rest. Roughly speaking, the top four pitchers recovered from exhaustion according to the following schedule:

Day 0 (day of game): 0% rested
Day 1: 15% rested
Day 2: 50% rested
Day 3: 75% rested
Day 4: 100% rested

As you can see, by day four (four days after having pitched), a starter is fully rested. But even though the rotation mode is set at "always start highest rested," it was rarely the case that a fully rested starter higher on the rotation would start with three days of rest when a fully rested starter with four or more days of rest was available, even if that starter was lower on the rotation. That was a puzzle, but I soon found the solution. The starting assignments are determined on the day before a game, and on that day a starter who will be 100% rested tomorrow is still only 75% rested today. I tested that theory with the Sox and other teams and found I could accurately predict the next day's starters 100% of the time.

That means that a four-man rotation should theoretically be possible with four high-stamina starters, even though the game logic might balk at it. The problem, though, would come with the brutal schedules that were common in that era of baseball. Double-headers, in particular, would wreak havoc on a four-man staff. The only times that my Sox pitchers started with less than four days of rest were the result of double-headers. The solution to the problem, then, is the one that Kid Gleason and other managers at that time made: rely on the top three or four pitchers for the majority of starts and give a few starts here and there as necessary to the guys at the end of the bench.

And that's where reality runs up against the OOTP game logic. If OOTP determined the starting pitcher on the morning of the game rather than the night before, high-stamina pitchers like the White Sox top quartet would be starting regularly on three days of rest. As a result, the majority of starts would go to the top pitchers, which was historically the case, rather than being spread out more evenly among five or six starters as OOTP does it. That one change to the game logic would go a ways to making pitching stats more accurate for the pre-reliever era. There would still be the problems, such as the over-use of relievers, which are discussed more extensively in my previous posts, but at least it would be a start.

Last edited by joefromchicago; 10-15-2019 at 02:21 AM.
joefromchicago is offline   Reply With Quote
Reply

Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -4. The time now is 11:33 AM.

 

Major League and Minor League Baseball trademarks and copyrights are used with permission of Major League Baseball. Visit MLB.com and MiLB.com.

Officially Licensed Product – MLB Players, Inc.

Out of the Park Baseball is a registered trademark of Out of the Park Developments GmbH & Co. KG

Google Play is a trademark of Google Inc.

Apple, iPhone, iPod touch and iPad are trademarks of Apple Inc., registered in the U.S. and other countries.

COPYRIGHT © 2023 OUT OF THE PARK DEVELOPMENTS. ALL RIGHTS RESERVED.

 

Powered by vBulletin® Version 3.8.10
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright © 2020 Out of the Park Developments