Distribution of performance

jcard · 03-26-2024, 11:08 AM

Does anyone have observations on how the distribution of statistics in v25 compare to real-life norms. In limited experience on prior versions, I found the curve to be noticeably flatter than in reality. Looking at the Stats+ pages of online leagues almost universally reinforces this perception. The issue is not so much extreme outliers as far too many players being at what are normally the far tails of the real MLB curves. Using the last league that I studied as an example, the highest wRC+ among qualifiers was 200 (higher than a typical season leader, but also not unprecedented for any given year) while the lowest was 25 (no way you get enough rope to be a qualified hitter with (non)production like that). More concerning, simply because it involves a greater percentage of the players than the outlier cases, is the number of qualified batters with wRC+ either below 60 or above 160 (just to use rough markers of, respectively, non-MLBer and MVP candidate calibers). Similarly, the 95-110 range is underrepresented.

RonCo · 03-26-2024, 03:31 PM

In the case of online leagues, which is all I can really speak for these days, it's hard to break out the effect of a non-focused GM. Or a GM who is fine with not losing, and doesn't have to suffer the fate of getting fired. Or...

jcard · 03-26-2024, 09:19 PM

Quote:

Originally Posted by RonCo

In the case of online leagues, which is all I can really speak for these days, it's hard to break out the effect of a non-focused GM. Or a GM who is fine with not losing, and doesn't have to suffer the fate of getting fired. Or...

I agree, though to see the pattern repeated among a majority of the leagues made the likelihood it was systemic in nature at least somewhat stronger. Another complication is fact neither OOTP nor Stats+ adjusts wRC+ for ballpark (whereas OPS+ is neutralized). As might be expected, catchers and middle infielders dominate the left tail. The production is so low, however, that if defense is the justification for keeping these players in the lineup then defense at a key position is being overvalued either in terms of the perception of AI managers or by the engine itself.

Thanks for the response. I recognize your name from when I first came upon OOTP a little over a decade ago, when I believe you were a more frequent presence here. If I recall, you had contributed some empirical studies of the game that I found quite informative and enjoyable.

RonCo · 03-27-2024, 05:07 PM

Thanks!

Your views of the lower tiers of leagues being off was something I was looking at some years back. I'm sure the AI did some weird valuations st times, but all my real interest is in paying against humans, so that's where I focus these days. And the social aspect of the online game does change output. On one hand, you have arguably three most intensely competitive set of OOTP GMs on b the planet, and on the other you have real life as it crashes into the sim schedule.

03-26-2024, 11:08 AM	#1
jcard All Star Reserve Join Date: Oct 2015 Posts: 610 Infractions: 0/1 (1)	Distribution of performance Does anyone have observations on how the distribution of statistics in v25 compare to real-life norms. In limited experience on prior versions, I found the curve to be noticeably flatter than in reality. Looking at the Stats+ pages of online leagues almost universally reinforces this perception. The issue is not so much extreme outliers as far too many players being at what are normally the far tails of the real MLB curves. Using the last league that I studied as an example, the highest wRC+ among qualifiers was 200 (higher than a typical season leader, but also not unprecedented for any given year) while the lowest was 25 (no way you get enough rope to be a qualified hitter with (non)production like that). More concerning, simply because it involves a greater percentage of the players than the outlier cases, is the number of qualified batters with wRC+ either below 60 or above 160 (just to use rough markers of, respectively, non-MLBer and MVP candidate calibers). Similarly, the 95-110 range is underrepresented.

03-26-2024, 03:31 PM	#2
RonCo Hall Of Famer Join Date: Aug 2003 Posts: 10,502	In the case of online leagues, which is all I can really speak for these days, it's hard to break out the effect of a non-focused GM. Or a GM who is fine with not losing, and doesn't have to suffer the fate of getting fired. Or...

03-27-2024, 05:07 PM	#4
RonCo Hall Of Famer Join Date: Aug 2003 Posts: 10,502	Thanks! Your views of the lower tiers of leagues being off was something I was looking at some years back. I'm sure the AI did some weird valuations st times, but all my real interest is in paying against humans, so that's where I focus these days. And the social aspect of the online game does change output. On one hand, you have arguably three most intensely competitive set of OOTP GMs on b the planet, and on the other you have real life as it crashes into the sim schedule.