alternate historical question

[Brad K]

First analysis. The raw data may look about the same at a glance to first sample but it isn't.

[Brad K]

Hey hef, need the game numbers for Ray Jablonski to complete this analysis. Totals in chart include his RL performance.

Unlike the earlier samples this one does not support the idea the best players would hit more HRs in your sim.

[thehef]

Originally Posted by Brad K

Hey hef, need the game numbers for Ray Jablonski to complete this analysis. Totals in chart include his RL performance.

You've got it right for Jablonski. I guess because OOTP favored Vern Benson (who played in the minors in '53 IRL) at 3B for the Cards, Joblonski spent '53-OOTP in AAA.

[thehef]

Originally Posted by Brad K

Unlike the earlier samples this one does not support the idea the best players would hit more HRs in your sim.

What may or may not explain it is that in my earlier test I did more forcing of the actual rosters for the four PCL teams that I plugged into the MLB. For this most recent test and changed some settings to allow AI to make roster moves. So it's possible that those four "former"-PCL teams had somewhat better pitching... I dunno.

[Brad K]

Is something different in the samples? Is the comparing the top 25 in the game against their RL performance different than comparing the top 25 in RL against their game performance?

It jumps out at me that in the top 25 RL comparison the total ABs is very close while it is not in top 25 game. What does this mean? Does it mean anything?

[thehef]

Originally Posted by Brad K

Is something different in the samples? Is the comparing the top 25 in the game against their RL performance different than comparing the top 25 in RL against their game performance?

It jumps out at me that in the top 25 RL comparison the total ABs is very close while it is not in top 25 game. What does this mean? Does it mean anything?

Hey Brad, apologies but I'm not understanding this. Can you please re-phrase the questions, etc.? And I'll try my best to answer

[Brad K]

We wanted to see if the best HRers did better in your sim with four weak teams than they did in real life. There were two different ways of determining the top HR hitters.

One was to use the top 25ish in the game and compare them to their RL performances. Another was to use the top 25ish in real life and compare them with their game performances.

The first sample method resulted in better performance in the game. The second sample did not. But we estimated performance in the game should be better than RL which would lead to the idea there might be something wrong with the method used for the second sample.

[Brad K]

We wanted to see if the best HRers did better in your sim with four weak teams than they did in real life. There were two different ways of determining the top HR hitters.

One was to use the top 25ish in the game and compare them to their RL performances. Another was to use the top 25ish in real life and compare them with their game performances.

The first sample method resulted in better performance in the game. The second sample did not. But we estimated performance in the game should be better than RL which would lead to the idea there might be something wrong with the method used for the second sample.

[Brad K]

Here's a different sample. We did best in game vs their real life performances. We did best in RL vs their game performances. In both cases players were compared against themselves.

Here's a comparison of the best in the game vs the best in RL Players are not the same.

[thehef]

So in reviewing each analysis of the results and going back to my original questions - "For best results, do I need to do anything with League Totals, Modifiers, the era_stats or era_modifiers files, etc.? Or just roll with it?" - I think the answer are "no" and "just roll with it."

In looking at HR rates and HR rates per 550 AB's, they seem really close. The slight differences may be attributable to the fact that batters were facing a bit poorer pitching overall (due to the presence of four teams with primarily AAA-level pitching), but they are not so different to suggest that anything would need to be done with those various files. At least that's my conclusion based upon this limited testing. Agree?

[Brad K]

The slight differences aren't attributable to the batters or pitchers. They're due to variance from the game's attempt to create historical output. Regardless of the actual talent in the league the league totals file will try to hammer the league into historical per 550 performance and aside from normal randomness or the talent in the league changing during the season (injuries) will be highly successful.

If, in your case, the additional players change the comparative talent of hitting and pitching/defense, then adjusting the modifier file is appropriate. Such adjustments, done accurately, would result in the league overall performing "as it should" but I don't know if the same is true of individual players.