Opening Day Dice Roll?

[zrog2000]

Originally Posted by bobbycockstrong

All else equal, if some mike trouts have a great season then the game will demand some of the other mike trouts have a subpar season to balance things out.

No, that's not what is demanded. If some Mike Trouts have a better than expected season, then the game will demand that some of the other players in the league have a subpar season to balance things out. There are no adjustments made for each individual Mike Trout. If it works out that way, it's coincidence, not mandated. It's a coincidence that will happen often, but it is not demanded by the game.

[bobbycockstrong]

Originally Posted by HRBaker

We have discussed what might happen if the season tracks too high or too low - maybe there is an internal "adjustment" across all players to pull things back to center - but I honestly don't believe even this adjustment is there - because it falsifies randomness and the law of averages.

Yeah but what im getting at is that its not pure "randomness" - its confined by restrictions (annual league totals). Ive mentioned annual league totals as well, but surely there are individual player ranges as well. If it was pure randomness, that would mean that things could happen like a player hitting 100 HRs or someone bats .600, because every AB would be purely random, but its not its confined by a system to keep the "randomness" in a specific range so its not truly random every AB. That would be problematic because it would not represent "realistic" baseball and stuff would be all over the place.

We dont know how the game handles this or corrects things, and maybe we never will (nor do we really need to). The only thing we know based on people looking at end of season totals across leagues, is that the game demands certain annual league totals fall into a certain range.

Anyways, we will likely never know, and it doesn't even matter really at the end of the day (I just personally find it interesting how the game handles these things, hence the back and forth).

[bobbycockstrong]

Originally Posted by zrog2000

No, that's not what is demanded. If some Mike Trouts have a better than expected season, then the game will demand that some of the other players in the league have a subpar season to balance things out. There are no adjustments made for each individual Mike Trout. If it works out that way, it's coincidence, not mandated. It's a coincidence that will happen often, but it is not demanded by the game.

Yeah i get that, see my other post below. I meant that in an all else equal comparison of the mike trouts. And again, im not saying the game "targets" individuals specifically, im saying that if some individuals perform above the baseline, then the game will demand that others perform below it to keep things in balance - so individual players are indirectly kept down to balance things out because there comes a point where they cant also perform above average because enough players already have. I just keep using the mike trout example because it came up here and also in a few other threads.

[The_Savage_1]

Originally Posted by bobbycockstrong

Yeah but what im getting at is that its not pure "randomness" - its confined by restrictions (annual league totals). Ive mentioned annual league totals as well, but surely there are individual player ranges as well. If it was pure randomness, that would mean that things could happen like a player hitting 100 HRs or someone bats .600, because every AB would be purely random, but its not its confined by a system to keep the "randomness" in a specific range so its not truly random every AB. That would be problematic because it would not represent "realistic" baseball and stuff would be all over the place.

We dont know how the game handles this or corrects things, and maybe we never will (nor do we really need to). The only thing we know based on people looking at end of season totals across leagues, is that the game demands certain annual league totals fall into a certain range.

Anyways, we will likely never know, and it doesn't even matter really at the end of the day (I just personally find it interesting how the game handles these things, hence the back and forth).

No, the simple mathematics of probability ensures this never happens, and by never I don't literally mean never, I mean some percentage starting with 0. followed by numerous zeroes.

[bobbycockstrong]

Originally Posted by The_Savage_1

No, the simple mathematics of probability ensures this never happens, and by never I don't literally mean never, I mean some percentage starting with 0. followed by numerous zeroes.

It was my understanding that the values were restricted by a baseline based on the league year that the simulation was based on (this seems generally accepted around here). There are literally options in OOTP single player to modify this stuff (so its not a secret). And some folks claim that the PT world was based on 2010 or 2011 statistic baselines (so even though OOTP single player is different, the same principle applies to PT apparently).

Therefore its not pure randomness, its randomness within a predetermined range and which is also modified by the output of other players (so that predetermined range is perhaps dynamic across the 162 games to keep the values "stable"?). How can something be truly random if its output is corrected by something else (the baseline annual league totals)?

Thats kind of all im saying. That its not truly random because the game make corrections (and thats fine). Statistics and player output are modified by the game ("the game demands it"), but i havent and am not suggesting that it targets specific players - just that as long as some players "overperform", it will happen that the game will demand specific players will "underperform" to achieve the desired annual league totals.

[HRBaker]

Originally Posted by bobbycockstrong

It was my understanding that the values were restricted by a baseline based on the league year that the simulation was based on (this seems generally accepted around here). There are literally options in OOTP single player to modify this stuff (so its not a secret). And some folks claim that the PT world was based on 2010 or 2011 statistic baselines (so even though OOTP single player is different, the same principle applies to PT apparently).

Therefore its not pure randomness, its randomness within a predetermined range and which is also modified by the output of other players (so that predetermined range is perhaps dynamic across the 162 games to keep the values "stable"?). How can something be truly random if its output is corrected by something else (the baseline annual league totals)?

Thats kind of all im saying. That its not truly random because the game make corrections (and thats fine). Statistics and player output are modified by the game ("the game demands it"), but i havent and am not suggesting that it targets specific players - just that as long as some players "overperform", it will happen that the game will demand specific players will "underperform" to achieve the desired annual league totals.

I think why the hairs on the back of our necks keep rising is the wording you insist on using; "the game will demand specific players will "underperform". That simply does not happen. "IF" there is an internal adjustment, it applies to "everyone" - even the guys that are over performing. Any other way makes the game a farce. Adjusting everyone down ""2%"" still accomplishes the goal of targeting the league totals without picking anyone out specifically.

[max venerabel]

Mostly what I've gotten from this conversation is that there are two competing concepts:



1) The devs say that there are no slump mechanics, and it would make the game unrealistic if there were, so there must not be any.


2) The observational data seems to bear out a greater than expected divergence of performance between identical rated players.


I don't think either concept #1 or #2 has been proven definitively, though both have been stated.



For #2, one attempt at an OOTP league with as few variables as possible has been made, and the reported data does show very wide swings in BABIP (as well as other stats) for identical players over a one season sample, which seems somewhat surprising.

In Perfect Team, which we know has certain differences from standard OOTP settings (such as increased player fatigue), it would be difficult to test this due to the many chaotic variables in PT leagues, i.e. varied divisional competition, park factors, etc. However, observationally, a surprising delta is present season to season..



For #1, I haven't yet seen a link where a dev definitively states there is no mechanism that could cause a good year/bad year assessment on players. This may be present, and perhaps I just overlooked it, but so far no developer interaction has occurred in this thread.

I also don't think this mechanic would necessarily make the sim "artificial". PT developers knew that there would be a high instance of identical players on teams, and perhaps they made the call that it would make game-play better if they didn't all perform too closely to their assumed ratings year to year, and added some randomness in service of this. In standard OOTP leagues, player development, aging and injuries add to the realism of fluctuation in performance over time, as well as to the depth of game play, so I could see this decision having been made.

[HRBaker]

Here...

[chazzycat]

Originally Posted by max venerabel

For #2, one attempt at an OOTP league with as few variables as possible has been made, and the reported data does show very wide swings in BABIP (as well as other stats) for identical players over a one season sample, which seems somewhat surprising.

Any surprise here is due solely to lack of understanding, IMO.


There is a huge amount of random variation in baseball.

[bailey]

If the game has pre-determined league totals within a certain % variance, then with enough data, you could determine if the last 2 weeks were going to be low-scoring or high-scoring and go into the AH and buy the appropriate stud if you were in a tight pennant race?

[zrog2000]

Originally Posted by chazzycat

Any surprise here is due solely to lack of understanding, IMO.


There is a huge amount of random variation in baseball.

That's what so many baseball fans today fail to realize. I can't tell you how many gameday threads I've been in where people start bitching about their favorite team by the 3rd inning because they haven't gotten a hit yet even if they're hitting the ball hard constantly.

The goal in baseball for hitters is to hit the ball hard. The goal is not to get weak hits because there is no skill to getting weak hits.

The goal in baseball for pitchers is to induce weak contact. It is not to give up line drives hit directly at people because there is no skill in getting loud outs.

[HRBaker]

Ugh. For some of these points we go round and round and never come to a conclusion. Let's try and very simple example...

(1) It's decided to match the league stats of a particular year, and the BA of that year (for the league) is .280.

(2) League variables are set to match .280 and the season begins.

(3) IF there are no further adjustments made, and IF the formulas within the game are stable, the final BA for the league replay might be .285 or might be .275. Players in the league may range from .350(MVP) to .000.

Before I go on to point #4, THIS (#3) is how I think OOTP works.

(4) However... IF there is a "mid course correction" as some here want to believe, the season is maybe 1/3rd done, and the game checks where we are. A .280 BA was the target, but the league is currently at .300, and the game determines it needs to slow things down.

(5) IF #4 is true, then the game probably sets the new target BA at .270, and ALL players have now been pulled down a bit by limiting their upper possibilities.

The guy leading the league with a .360 BA may still win the Batting award, or he may be overtaken by someone else. Each player still has a full range of results - except the top end has been clipped a bit.

I do not think OOTP works this way, but to answer someone's question above; YES, if you knew what the adjustments were, you would be able to take advantage of it - BUT your obviously never going to know what they were because OOTP wouldn't tell you.

Matt Arnold has said there are no such adjustments concerning players, and If there WERE such adjustments for the league - that wouldn't effect player stat variances that were mentioned here to start the discussion.

[max venerabel]

I had another thought on this.

If you look at the chart of Mike Trouts posted earlier in this thread, what stands out is the 129 point swing between the best BABIP season posted (.402) and the worst (.272). That seems pretty wild for 24 identical players.

However, if you look at MLB stats for 2018, the performance of the 24 Trouts are far off from the general range of BABIP for qualified hitters over the course of the season. The top BABIP was .375, the bottom was .217, for a 158 point swing.

Perhaps OOTP simply looks at all hitters (and maybe pitchers) as completely equal in terms of BABIP, so it doesn't matter if they are 24 identical players or 24 random players.

The real Mike Trout, for example, seems to have a more steady year-to-year BABIP than a sample of 24 random MLB players would (.318 in his worst full season to .383 in his best, for a swing of 65 points). I'm not certain what the typical bell curve of a single player's career is like in terms of BABIP, but if the curve is softer than for players overall, yet PT doesn't (for whatever reason) model that curve in its mechanics, perhaps that's all that's going on.

[zrog2000]

Keep on expecting cards in PT to perform like players IRL and you're never going to be satisfied and complain about it forever.

[max venerabel]

Originally Posted by zrog2000

Keep on expecting cards in PT to perform like players IRL and you're never going to be satisfied and complain about it forever.

So wait a moment, on one had we are encouraged to consider the absolute purity of modeling in OOTP to be so complete that questioning its perfection is tantamount to heresy....


... and on the other hand we should never expect any correlation to real world outcomes.

[zrog2000]

Originally Posted by max venerabel

So wait a moment, on one had we are encouraged to consider the absolute purity of modeling in OOTP to be so complete that questioning its perfection is tantamount to heresy....


... and on the other hand we should never expect any correlation to real world outcomes.

Yes. This is what real baseball would look like if the talent level was about 100 times better than it is now.

It wouldn't be shocking if Mike Trout struck out 4 times in the All-Star game, just like it wouldn't be shocking if Jacob deGrom gave up 4 home runs in a row. Or if Trout hit 4 HR and deGrom struck everyone out. And imagine the variability if there were 162 All Star games played.

And the All-Star game isn't even close to the talent level of PT perfect leagues.

[max venerabel]

Originally Posted by zrog2000

Yes. This is what real baseball would look like if the talent level was about 100 times better than it is now.

It wouldn't be shocking if Mike Trout struck out 4 times in the All-Star game, just like it wouldn't be shocking if Jacob deGrom gave up 4 home runs in a row. Or if Trout hit 4 HR and deGrom struck everyone out. And imagine the variability if there were 162 All Star games played.

And the All-Star game isn't even close to the talent level of PT perfect leagues.

I really don't see how you can present a correlation between outcomes in a single ASG v.s. ten seasons of play for 24 identical players.


That's a difference of sampling 5 plate appearances vs. like 170,000.


I don't think better competition should mean wider swings when sample sizes are that large.

[zrog2000]

Originally Posted by max venerabel

I really don't see how you can present a correlation between outcomes in a single ASG v.s. ten seasons of play for 24 identical players.

That's a difference of sampling 5 plate appearances vs. like 170,000.

I don't think better competition should mean wider swings when sample sizes are that large.

It's not that and it was only an example, not proof using sample sizes. If Mike Trout faces Jacob deGrom for 600 PA, what are the likely outcomes? I'd say that there's an absolutely huge range, way bigger than if Mike Trout gets 600 PA against pitchers that barely belong in the majors. The much higher talent level for all players in perfect leagues makes the variation much higher. You can understand that if you think about my example. The best batters of all time could either continue hitting the same way they always did or they could be completely dominated by Pedro Martinez along with everywhere in between.

I only stated that example because the All Star game is the only thing you ever see in real life that has a higher talent level then regular MLB baseball and PT perfect leagues have a way higher talent level than an All Star game.

If you can't understand it at all, then just keep on complaining forever.

[<Pion>]

Originally Posted by bobbycockstrong

It was my understanding that the values were restricted by a baseline based on the league year that the simulation was based on (this seems generally accepted around here). There are literally options in OOTP single player to modify this stuff (so its not a secret). And some folks claim that the PT world was based on 2010 or 2011 statistic baselines (so even though OOTP single player is different, the same principle applies to PT apparently).

Therefore its not pure randomness, its randomness within a predetermined range and which is also modified by the output of other players (so that predetermined range is perhaps dynamic across the 162 games to keep the values "stable"?). How can something be truly random if its output is corrected by something else (the baseline annual league totals)?

Thats kind of all im saying. That its not truly random because the game make corrections (and thats fine). Statistics and player output are modified by the game ("the game demands it"), but i havent and am not suggesting that it targets specific players - just that as long as some players "overperform", it will happen that the game will demand specific players will "underperform" to achieve the desired annual league totals.

Unless OOTP has gone through a massive change we don't know about, it doesn't work that way. When modifiers were introduced back in the early OOTP, it gave the player a way to try to get overall statistics that matched the season the OOTP player was simming. Without modifiers, the game was totally at the mercy of the ratings of the ballplayers to determine stats. That would results in unrealistic historical sims. By adjusting the modifiers, the number of HRs or the batting averages of the players as a whole for the season would more closely follow what happened in real life. It made it closer, it did not make it exact. Players did not stop hitting home runs if the total home runs wished for was reached and modifiers did not change to attempt to squash home runs in the less the person running the league manually changed them. In one of the releases, modifiers were automated so that historical leagues would be able to have the more realistic results without needing a 3rd party tool to tell them how to change the modifiers.

PT is not historical, so I don't think we know exactly how and or if the modifiers change from year to year. I believe they do to get some variances in seasons outside of what pure randomness would produce, but I don't know if the Devs have ever confirmed that or if anyone has studied year to year variances to determine the likelihood of modifier change. I don't believe they change during the year or that for each result for player A, one of the other players is going to automatically have less to offset it. As someone mentioned, mathematical randomness/probability takes care of it for us within the framework of the dance between modifiers and the ratings of the players in the individual league getting the most playing time.

The players playing should have an important part. For example, regardless of how high the league is (Diamond versus Bronze), the league stats tend to remain about the same. That's because overall, the interaction between pitchers and hitters remains fairly constant against each other (as hitter ability increases at higher level, pitching does at well). But if a Diamond league hits a certain number of HRs in a season, had that same season been played with the same pitchers but all iron players, they would not have the same number of HRs. Unless things have changed, the game would not make adjustments to try to reproduce the same number of HRs had the stars played instead of irons. Same would be true if all great hitters were used like normal but everyone pitched the worst players they could find. HRs would balloon greatly rather than self adjust.

I could be wrong, but that's the way it worked in the past.

[HRBaker]

Originally Posted by <Pion>

...PT is not historical...

Excellent point. That means OOTP doesn't - and can't - nor do we want it to - replicate past stats exactly. It is meant to be "close" and "close" is defined multiple ways by multiple people.

For me, "close" is close enough.