Six Simulations to Test LHP and RHH

[grmagne]

This is a follow-up to a previous thread where I asked whether there are too few left-handed in the OOTP major leagues.There are similar things going on with right-handed batters and I’ll describe both in depth below. Previous thread:

http://www.ootpdevelopments.com/board/ootp-17-general-discussions/265826-vanishing-left-handed-pitchers.html

My computer was hard at work this past week running six simulations. Each simulation used “New Standard Game” from the main menu with leagues running from 2016-2080. The 2035-2080 league data was then exported Microsoft Access and Excel. The only changes I made to the default settings were removing League Evolution plus these scenario-specific changes:

1. Default Settings: This is the baseline scenario.
2. Sabermetrics: The AI used Sabermetric rather than Traditional lineups in all leagues and minor-league managers pinch-hit “Very Often” rather than “Rarely”.
3. No Scouting: The scouting system was disabled.
4. (a) Modified Development: Pitcher Aging Speed = 0.5, Pitcher Development Speed = 1.5, Batter Aging Speed= 1.4, Batter Development Speed = 0.6
   
   (b) Modified Development:P itcher Aging Speed = 1.5, Pitcher Development Speed = 1.5, Batter Aging Speed= 0.6, Batter Development Speed = 0.6
   
   (c) Modified Development: Pitcher Aging Speed = 0.6, Pitcher Development Speed =0.6, Batter Aging Speed =1.5, Batter Development Speed = 1.5

The purpose of scenarios 2 & 3 was to test whether the AI would promote LHP through the minors more frequently if (2) the RHP platoon advantages over LHP were minimized a bit or (3) if LHP and RHP would be promoted more equally with perfect knowledge of their true abilities. The purpose of the various scenarios #4 was to test whether accelerating/decelerating pitcher & batter development affects the overall numbers of LHP and RHH in the major leagues.

SUMMARY:

First of all, virtually all of the OOTP data that I analyzed very closely matches the real statistics. Out of the Park is truly amazing in this regard! I’ll post a summary of my findings below to reiterate a couple of issues that are a bit off from MLB, and then follow-up with numerous charts so that you can see for yourself.

• There is definitely something causing left-handed pitchers and right-handed batters to get cut from the minor leagues more often than right-handed pitchers and left-handed batters. They both disappear throughout the minor league systems, though in many simulations they make a slight comeback in the majors.
• In most simulations right-handed hitters had longer careers than left-handed hitters. RHH also have a noticeably higher average age than LHH in all my simulations. In the MLB from 1970-2015, RHH were only 0.1 years older on average than LHH, which is statistically insignificant.
• Interestingly, LHH start tobecome just as prevalent as RHH at around age 40 in real MLB, with a noticeable increase in LHH% between ages 35-40. The older age data samples are smaller so it’s difficult to know if this is a fluke or a real trend that will endure into the future.
• Starting LHP were only 20.7% -25.3% of my OOTP major league simulations versus 29.8% in real MLB, 1970-2015. So there’s definitely an underrepresentation there. LHP were over 29% of draftees in my simulations so it’s difficult to understand why so few make itto the majors.
• Relief LHP are much better represented in OOTP. They were 24.2% - 28.0% of major league LHP in my simulations versus about 30.6% in MLB. Pretty close.
• Right-handed batters are 51.5%- 57.6% of major league batters versus 54.2% in real MLB, 1970-2015. Perfect!
• The strange thing in the point above is that right-handed batters are about 67% of draft classes but rapidly disappear from the minor leagues before landing in the “correct” range at the major league level. I have no idea why this happens.
• Simulation #2, Sabermetrics, wasn’t as effective as I had hoped because even with the Sabermetric lineup selection, minor league managers only slightly vary their lineups according to the opposing pitcher’s throwing hand. Often it’s the same starting players against both, except the LHH move down the lineup against LHP. On occasion there was an actual platoon in the minors.
• Simulation #3, which removed the scouting system from the game, had the highest proportion of LHP and RHH making it to the majors of all the simulations. My interpretation is that inaccurate scouting and poor performances hold back LHP and RHH in other simulations, but that is partially corrected here. And yet, the age gap between LHB and RHB was at its greatest in this simulation, meaning that slower aging of RHH is likely a real part of the code.
• Simulations 4a and 4b, which both increase pitcher development speed, had fewer LH starting pitchers in the majors but more LH relievers. That was the opposite of what I had predicted.
• Simulation 4c, which reversed the development & aging assumptions of simulation 4b, had the lowest percentage of LH Relief Pitchers and highest percentage of RH batters in the major leagues.

[grmagne]

LH Starting Pitchers are disappearing from the minors in every simulation, and always underrepresent at the major league level, though different settings will have some effect on this.

[grmagne]

LH Relief Pitchers make it to the major leagues at a much better rate than LH Starters. But there’s still a slight underrepresentation at each age, as later charts will show.

[grmagne]

Right-handed batters seem to have even higher minor league attrition rates than LHP but they end up having roughly the correct percentage in the major leagues (compared to MLB).

[grmagne]

These charts show the average ages by the types of major league players. Since left-handers and right-handers seem to develop at different speeds in this game, I just wanted to take a high-level snapshot of all scenarios. These ages are all weighted by playing time: for pitchers it’s weighted by number of starts and number of relief appearances, while batters are weighted by at bats.
Left- and right-handers in real MLB seem to have similar career lengths and average ages but left-handed relievers have the longest careers and highest average age of any group.

[grmagne]

Here are the percentages of major league left-handed starting pitchers by age, comparing real MLB versus simulation. As you can see, there are far fewer left-handed pitchers in OOTP than real MLB. However, most (but not all) simulations show an increasing number of LHP at older ages.

[grmagne]

Similar chart, this time for left-handed relievers. For reasons I'm not sure of, LHP are more likely to make it the majors in OOTP if they're a reliever rather than a starter. Or maybe LHP are more likely to be moved to the bullpen upon reaching the majors?


In both real MLB and OOTP, left-handed relievers have longer careers than right-handed relievers but the difference in average age is larger in real MLB.

[grmagne]

Here are the percentages of at bats by left-, right-, and switch-hitters. An amazing thing in real MLB is that between 1970 and 2015 left-handed hitters are almost equal in numbers to right-hand hitter sat older ages. There’s less data at older ages, of course, so I have no idea if that trend will continue into the future or whether it’s just a fluky result.

In many of these simulations the percentage of OOTP right-hander batters goes up significantly at older ages. I wonder if there something in the code that causes them to age more slowly than left-handed batters?

[grmagne]

These charts show the age and handedness distribution of all players in the OOTP simulations. Just in case you’re interested in how the initial settings affect age distributions.


This first chart is for starting pitchers.

[grmagne]

Same type of charts, but this time for batters.

[MrPhysics]

You have a lot of time on your hands. That said, I am impressed with the thorough analysis!

[Matt Arnold]

Thanks! That's a lot to take in. We certainly try to get it as accurate as possible, but I'm sure it would take a ton of studying to figure out, for example, why there's relatively speaking more LH bats at older ages than RH bats. Is it based on the type of player? Platoon advantages? Manager preferences? Or are they actually developing/aging differently?

Certainly interesting case studies.

[grmagne]

Originally Posted by Matt Arnold

Thanks! That's a lot to take in. We certainly try to get it as accurate as possible, but I'm sure it would take a ton of studying to figure out, for example, why there's relatively speaking more LH bats at older ages than RH bats. Is it based on the type of player? Platoon advantages? Manager preferences? Or are they actually developing/aging differently?

Certainly interesting case studies.

Thanks for the response. Don’t worry, I am consistently impressed at how closely OOTP matches MLB and I think the developers have all done an amazing job – and the game gets better every year! I wouldn’t have bothered doing all this analysis except for the fact that right-handed platoon hitters sit on the bench too much during a season and it would be nice (and more realistic) if there were more left-handed pitchers in the league. There isn’t a user option to increase the number of left-handed pitchers so instead I did a full analysis to share my observations with the developers in the hopes that someone can address it.


Honestly, it seems to me that there is an extremely simple fix for this, and I can’t imagine any side effects to other parts of the game from increasing the number of draft-eligible LHP….except that right-hand platoon hitters would get to play more.


Here are my suggestions and I really hope the solution is as simple as it looks:

1. It seems right now that newly generated fictional pitchers are about 29% left-handed and 71% right-handed. If you bumped that up to 35%/65% I think that would immediately solve the problem of too few left-handers in the majors. There'd be a few too many in Rookie ball, but I can't imagine anyone other than me would ever noticed 2%-3% too many LHP in the Rookie leagues.
2. If you're willing to do #1, I'd like to make a 2nd request: can you bump up the number of switch-hitters? Right now it looks like about 10% of newly generated fictional hitters are switch-hitters. Somewhere in the 12%-14% range would be more consistent with modern-day MLB (taking away a corresponding number from LHH & RHH in equal number would make sense).