Production drop-off in playoffs

[Brad K]

Originally Posted by Charlie Hough

without penalizing them or trying to prevent them from generating more of those results once their individual totals get to certain levels.

It's done in advance. Players with outlier seasons in real life for home run production have their HR rating reduced.

[Charlie Hough]

Originally Posted by Brad K

It's done in advance. Players with outlier seasons in real life for home run production have their HR rating reduced.

Show us the proof of this, and then have OOTP Developments show up here and confirm that it has coded the game to do this. When you look at cases such as Brady Anderson, who is probably the single biggest HR outlier in the history of baseball, he gets exactly the power rating you would expect.

In a 1996 OOTP game, he gets one of the top-four best power ratings in all of MLB. McGwire gets a 100 overall power rating on a 100-point scale because he hit 52 HR in just 548 PA and 423 AB. He is rightly #1. Anderson gets an 85 and is #4 in power, just 1 point behind Gonzalez and Griffey, because he hit his 50 HR in 687 PA and 587 AB. They hit 47 and 49 HR in fewer PA and AB, so they technically get better power, but only by a single point overall. Since Anderson's rating is almost identical to theirs overall, he should theoretically be able to hit something quite close to his 50 HR real life total, especially if he gets PA and AB similar to his real life totals. Based on these numbers, there is no way that OOTP has reduced Anderson's power rating as an outlier. It clearly doesn't do that with him.

Unless there are many outliers like Anderson who aren't calculated in the same way, which I haven't seen, then I'm not sure where you're getting this idea that players with outlier seasons for HR production are having their HR rating reduced.

Even if that were the case for certain HR outliers, it would be a narrow function of historical player ratings creation for those specific outliers. It definitely doesn't have anything to do with league totals, league total modifiers, league totals vs. player ratings in generating results, or adjusting the statistical probabilities of in-game outcomes based on previous simulated results, which is what we're discussing here.

[Brad K]

Originally Posted by Charlie Hough

Show us the proof of this, and then have OOTP Developments show up here and confirm that it has coded the game to do this.

So not only do I have to show it's true I have to have a dev show up and confirm what I've proven. Wow.

Anyway, Lukas showed up in a previous thread and confirmed it. In one of the threads below Syd commented on the dev's confirmation. So it's common knowledge among some people. Here's some reading for you.

https://forums.ootpdevelopments.com/...d.php?t=331438

https://forums.ootpdevelopments.com/...d.php?t=331958

https://forums.ootpdevelopments.com/...d.php?t=346122

There are more but this will make the issue clear.

[Charlie Hough]

I have seen two of those threads before, and the evidence is not compelling at all. As others pointed out in those other threads, there are many reasons why the observed outcomes probably happened, based on game settings, the type of stats being used, how statistical probabilities work, randomness, the conditions of your particular game, and other possible factors. There is nothing in any of those results that directly suggests or proves that certain outliers' power ratings or home run output are being nerfed in OOTP. Also, I don't see the post from Lukas, which you claim admits to deliberate nerfing of players' power ratings.

However, for some truly compelling evidence, let's take a look at my screenshot below. I just went ahead and simulated the full 1996 season featuring Brady Anderson. Take a look at the top home run hitters in MLB, and notice who hit 51 home runs in 583 AB, to almost identically match his real-life performance. I've already posted Anderson's power rating and the discussion of that rating relative to his stats and the other real-life leaders, showing that his power rating has not been reduced at all. Now you have the statistical output to match it.

This is just one example, but it's clear evidence that contradicts any claim that OOTP is nerfing the power rating or output of outliers such as Anderson. If OOTP Developments has actually coded the game to reduce certain outliers' home run output in the sim, then by your definition in another thread, clearly "it's not working."

[Lukas Berger]

I'm pretty sure I did say say that in the past, but if I recall correctly, I was specifically talking about Roger Maris in 61, not literally every outlier.

This also depends a little on how you define outlier. In the sense of McGwire or Maris, we do generally somewhat nerf their record setting outlier seasons when setting ratings, because those seasons if applied literally could easily result in 'unrealistic' totals a high percentage of the time, like Maris hitting 75 HR's.

But for those whose seasons are not historical league-wide outliers, but just personal outlier seasons, like Anderson, we wouldn't do anything specific to adjust those, beyond whatever the natural league and park adjustments they'd get.

[Brad K]

Originally Posted by Charlie Hough

I have seen two of those threads before, and the evidence is not compelling at all. As others pointed out in those other threads, there are many reasons why the observed outcomes probably happened, based on game settings, the type of stats being used, how statistical probabilities work, randomness, the conditions of your particular game, and other possible factors. There is nothing in any of those results that directly suggests or proves that certain outliers' power ratings or home run output are being nerfed in OOTP. Also, I don't see the post from Lukas, which you claim admits to deliberate nerfing of players' power ratings.

If the evidence isn't compelling then I can only conclude you didn't read the threads well. Your talk of performance in saves is additional evidence. The discussions weren't about performance in saves but rather ratings which, with a huge sample size, will eventually provide some signal amidst the noise in saves but aren't data in the sample size you've presented..

I have some reservations about posting some data as there is a possibility you're predisposed to say anything I post is not compelling. However I decided to post it as the audience, assuming there is one, might find it intersting.

Ratings are single year

Davey Johnson 1973 actual 43 42.8/550 rating 38/550

Brady Anderson 1996 actual 50 47.5/550 rating 45/550

Roger Maris 1961 actual 61 56.9/550 rating 45/550

Mickey Mantle 1961 actual 54 57.8/550 rating 50/550

Mark McGwire 1998 actual 70 75.6/550 rating 69/550

Mark McGwire 1999 actual 65 68.6/550 rating 64/550

Barry Bonds 2001 actual 73 84.3/550 rating 67/550


I actually don't have a problem with lopping off the peaks but I think it should also include filling in the valleys a bit. However my preference is for a floor and cap, both somewhat fuzzy.

[MathBandit]

Players who had outlier seasons irl *should* have their HR totals significantly lower in an accurate sim vs what they had in real life. That's what makes their season an outlier.

[Brad K]

Originally Posted by MathBandit

Players who had outlier seasons irl *should* have their HR totals significantly lower in an accurate sim vs what they had in real life. That's what makes their season an outlier.

Where is the line between skill and luck?

Should we give such a player some help in his bad luck seasons?

Not disputing what you say. Results in the game should be plausible.

Are we at some point calling real life implausible?

[Lukas Berger]

There are a lot of variables in play here determining the target rating beyond specifically intending to limit outliers.

Park factors, year and era, what amount of players you choose to 'make worse' or 'make bad' all factor into the target ratings.

[David Watts]

Originally Posted by MathBandit

Players who had outlier seasons irl *should* have their HR totals significantly lower in an accurate sim vs what they had in real life. That's what makes their season an outlier.

What about in a replay? If I have OOTP set to use real lineups/real transactions. Use 1 year recalc with development off. Turn coaches and scouting off. Should Maris still have his HR totals significantly lower? What about a guy like Norm Cash? 61 was a crazy outlier season for him. What's funny is, whenever I replay 61, Jim Gentile always seems to reach at least 50 home runs and I think I've seen him hit as many as 56.

[Brad K]

Originally Posted by David Watts

What about in a replay? If I have OOTP set to use real lineups/real transactions. Use 1 year recalc with development off. Turn coaches and scouting off. Should Maris still have his HR totals significantly lower? What about a guy like Norm Cash? 61 was a crazy outlier season for him. What's funny is, whenever I replay 61, Jim Gentile always seems to reach at least 50 home runs and I think I've seen him hit as many as 56.

Single season replay with historical lineups is a bit of a different game. OOTP acknowledges this with its recommendation to reduce the number of ABs to which adjust/weaken applies. Thus it seems a reduction in outlier HR ratings is less justified.

However there is something else. With every player playing his games in his historical parks, shouldn't all park factors be set to 1.000? I'm unclear if there's any adjustment to real stats for in game park factors but I lean toward there being none.

[Charlie Hough]

Originally Posted by Brad K

I have some reservations about posting some data as there is a possibility you're predisposed to say anything I post is not compelling.

I read those threads entirely and already looked at all the data that was posted, and none of it was compelling. It's not a matter of predisposition. It's a matter of what was posted. Lukas already replied and clarified what he originally said and what OOTP is doing, which generally backs up what I posted here and what I have seen OOTP do for many years.

Maybe in a very few select cases, involving specific individual outliers that we can almost count on one hand, one player rating calculation is adjusted in order to avoid potentially extreme outcomes at the far end of statistical possibilities for one specific stat. But that is an extremely narrow and specific adjustment that's designed to avoid a potentially egregious lack of realism in a high-profile stat. It has nothing to do with what was being discussed here, which is the fact that some baseball sims penalize players and start reducing the statistical possibilities of generating certain future outcomes based on previously simulated stats, and they do it in a generalized way, across all players, all in an attempt to enforce so-called realism at the expense of real statistical probabilities.

OOTP does not do that, and what it's doing with a few HR outliers in advance, by adjusting how their power rating is calculated, doesn't contradict that fact. It's also so limited to such a miniscule number of players and a single type of rating that it's inconsequential for literally every other player in the database and all other calculated ratings.

Now, maybe some people find it interesting or compelling anyway. But I don't, and it seems to make perfect sense to me, so that's the last thing I'm going to post on this topic.

[David Watts]

Originally Posted by Brad K

Single season replay with historical lineups is a bit of a different game. OOTP acknowledges this with its recommendation to reduce the number of ABs to which adjust/weaken applies. Thus it seems a reduction in outlier HR ratings is less justified.

However there is something else. With every player playing his games in his historical parks, shouldn't all park factors be set to 1.000? I'm unclear if there's any adjustment to real stats for in game park factors but I lean toward there being none.

I understand, I was just curious if OOTP was doing something wrong when it comes to single season replays? I've done them in the past and Maris will sometimes come very close to his real life 1961 statistics. Same goes for Cash. But, it seems that outliers should be falling way short of their real life statistics in order for a sim to be accurate.

[Lukas Berger]

Originally Posted by David Watts

But, it seems that outliers should be falling way short of their real life statistics in order for a sim to be accurate.

Not really, no.

Setting a target a bit short, just the way probability and statistical distribution works, outliers will still happen a good portion of the time. So there's always still a decent chance that he'll hit his real life numbers, but it's not necessarily the expected outcome in any given sim.

[David Watts]

Originally Posted by Lukas Berger

Not really, no.

Setting a target a bit short, just the way probability and statistical distribution works, outliers will still happen a good portion of the time. So there's always still a decent chance that he'll hit his real life numbers, but it's not necessarily the expected outcome in any given sim.

Yes, but Maris never hit more than 39 home runs in a season after the 1961 season. So, shouldn't OOTP take this into account for a 1961 single season replay to ensure that Maris hits significantly fewer home runs than the 61 he hit in real life? Shouldn't this be done the ensure a more accurate sim? Maybe a "bit" isn't enough.

[Brad K]

Originally Posted by Charlie Hough

I read those threads entirely and already looked at all the data that was posted, and none of it was compelling. It's not a matter of predisposition. It's a matter of what was posted. Lukas already replied and clarified what he originally said and what OOTP is doing, which generally backs up what I posted here and what I have seen OOTP do for many years.

My original point - that the HR ratings of high performers were reduced to lower the chance of them out performing outlier seasons - has been proven. Lukas' posts do not back up your position.

There's no reason for you to argue with me about this. Lukas is here. Tell him his game doesn't work the way he says it does.

[Lukas Berger]

Originally Posted by David Watts

Yes, but Maris never hit more than 39 home runs in a season after the 1961 season. So, shouldn't OOTP take this into account for a 1961 single season replay to ensure that Maris hits significantly fewer home runs than the 61 he hit in real life? Shouldn't this be done the ensure a more accurate sim? Maybe a "bit" isn't enough.

What Maris does the rest of his career, either before or after 1961, has nothing to do with anything.

'Outlier' in the sense I use it only refers to outlier results in the context of a given season, and has nothing at all to do with a player's career numbers or if their performance was a personal outlier.

The game is not adjusting Bartolo Colon's HR totals in 2016 down because his 1 HR was a personal outlier

[Brad K]

Originally Posted by Lukas Berger

There are a lot of variables in play here determining the target rating beyond specifically intending to limit outliers.

Park factors, year and era, what amount of players you choose to 'make worse' or 'make bad' all factor into the target ratings.

How do park factors interact with real stats?

[David Watts]

Originally Posted by Lukas Berger

What Maris does the rest of his career, either before or after 1961, has nothing to do with anything.

'Outlier' in the sense I use it only refers to outlier results in the context of a given season, and has nothing at all to do with a player's career numbers or if their performance was a personal outlier.

The game is not adjusting Bartolo Colon's HR totals in 2016 down because his 1 HR was a personal outlier

Lukas, I was just having some fun. Not being serious at all

[Lukas Berger]

Originally Posted by David Watts

Lukas, I was just having some fun. Not being serious at all

I suspected this after I responded and read your message again. Was having one of my overly serious days, I guess