Pitcher inconsistency

[HRBaker]

Originally Posted by Mizzery

...the concept that the cards are not performing in a static environment- each year, the competition level varies and will have a resultant impact on my player’s performance...

Another critical point often ignored. A few things we know...

1. OOTP does not "load" or "decrease" any players beyond their published stats. In other words, a "75" Contact rating is the same "75" for every other player who has a "75". Thus in the math world, ALL 75's have the same "hit" when you roll a "74". Accepting that premise, the only other variable from one player to the next or one team to another is the opponent's rating.

2. If there was a way to rate the strength of all teams in a league and calculate your average opponent's rating, you would find that each season brings a different value.

3. Add to that the schedule is different from year to year, and your 60HR Babe Ruth may face better pitchers one year than the next, and your variables slide even further.

4. and then there is just plain "luck" on your "dice roll".

What OOTP has made clear over the years is they don't "hide" any ratings. They don't add in anything for hot-and-cold streaks. They simply let the physics of "math" take care of all of that.

The result is no different that when my brother use to play a baseball game when we were kids using a deck of cards with singles, doubles, ground outs, etc. on them. It all depended on how you shuffled the deck. The only difference here is each player has a different deck.

I firmly believe that OOTP says there is nothing to fix because... there is nothing to fix.

[BPS]

Originally Posted by zrog2000

Just call out who you want to call out and stop with this cryptic trolling. I'm assuming you're talking about me while ignoring half of the post I made.

It is not cryptic at all.

Next up: writing "small sample size" is not always the universal kill shot many people think it is.

[zrog2000]

Originally Posted by BPS

It is not cryptic at all.

Next up: writing "small sample size" is not always the universal kill shot many people think it is.

Oh please impose more wisdom on we the stupid.

But instead of talking in obscure generalities, be specific.

[BPS]

Originally Posted by HRBaker

....They don't add in anything for hot-and-cold streaks...

My reading of what has been said is that if a player has been identified as being, say, "hot" is not "predictive" of future behavior. This is different than saying that no code creates hot players.

It very well might be true that the vast majority of those identified as hot are simply the beneficiaries of luck. But that does not mean others, perhaps small percentage, who are identified as hot actually have some factor that has made them hot and a chance exists that for some unpredictable time in the future this factor might remain operative.

Perhaps I'm reading the language that has been used more closely that is appropriate, but the language is all we got.

[zrog2000]

http://www.ootpdevelopments.com/boar...7&postcount=28

The higher variability comes because the talent ratings in most PT leagues are much more flat than a regular OOTP league. Even in PT though, you won't see Perfect Pedro or Perfect Babe having too many bad seasons, they're just too good even with the inflated talent pool. But for average to bad players, or even just good ones, a bad season is always possible both in regular OOTP and in PT.

It's no different than in regular OOTP, but the flatter talent pool and the big difference in who is a average and likely highly variable player makes things feel different and makes it hard to adjust mentally to who's that sort of player.

In regular OOTP, you wouldn't think twice if Anibal Sanchez or Mike Moustakas has a down year one year and a great year the next. Never mind actual bad players. You are not only not surprised, but expect it even.

In some PT leagues, because of the higher and flatter talent pool, Anibal Sanchez = Bob Gibson and Mike Moustakas = Brooks Robinson. That can be a hard thing to wrap your brain around, but the underlying mechanics and variance are the same in each case.

[HRBaker]

Originally Posted by BPS

My reading of what has been said is that if a player has been identified as being, say, "hot" is not "predictive" of future behavior. This is different than saying that no code creates hot players.

It very well might be true that the vast majority of those identified as hot are simply the beneficiaries of luck. But that does not mean others, perhaps small percentage, who are identified as hot actually have some factor that has made them hot and a chance exists that for some unpredictable time in the future this factor might remain operative.

Perhaps I'm reading the language that has been used more closely that is appropriate, but the language is all we got.

My meaning was that a player is "hot" because he HAD a series of good dice rolls. His current rating of "hot" can end the next 3 games by being 0-12 or posting a 6.50 ERA. EACH at bat and EACH inning pitched is determined whether it's in the middle of the season or the first game.

HOT is simply a notation of what has happened the past X number of games.

[BPS]

Originally Posted by zrog2000

Oh please impose more wisdom on we the stupid.

But instead of talking in obscure generalities, be specific.

They are not obscure at all.

If some effect appears when using a small sample, two possibilities exist: (1) the effect is real or (2) the effect is not real but only appeared due to chance. You cannot tell for sure which is the case. But, some folks assume that #2 is always the case.

Of course, it also is often the case that folks who think they have found some real effect (in a small sample) believe they have obvious evidence that the effect is real. This is equally incorrect.

[Orcin]

Here's what I know: some of my cards perform better than the equivalent card on other teams, and some of my cards perform worse than the equivalent card on other teams. This variation is consistent and occurs over a large sample size. There hasn't been any credible explanation for this variation other than dumb luck.

[CrazyWR]

ballpark factors, divisional opponents, quality of opponents, statistical variance.

[BPS]

Originally Posted by HRBaker

My meaning was that a player is "hot" because he HAD a series of good dice rolls. His current rating of "hot" can end the next 3 games by being 0-12 or posting a 6.50 ERA. EACH at bat and EACH inning pitched is determined whether it's in the middle of the season or the first game.

HOT is simply a notation of what has happened the past X number of games.

Yes, HOT is simply just the identification of past behavior (a series of good outcomes).

And, yes, outcomes in OOTP are determined by random variables. Yet, it is hypothetically the case that a series of past good outcomes can be partly due temporary boost in the probability of a good outcome.That is, some hotness could (hypothetically) be due to something other than pure randomness.

My guess is I'm wrong, but nothing the developers have said eliminates the above as a possibility.

[Orcin]

Originally Posted by CrazyWR

ballpark factors, divisional opponents, quality of opponents, statistical variance.

That’s what I said.... dumb luck. You can’t control any of these factors, so whether your card hits a string of favorable or unfavorable is pure luck.

[zrog2000]

Originally Posted by BPS

They are not obscure at all.

If some effect appears when using a small sample, two possibilities exist: (1) the effect is real or (2) the effect is not real but only appeared due to chance. You cannot tell for sure which is the case. But, some folks assume that #2 is always the case.

Of course, it also is often the case that folks who think they have found some real effect (in a small sample) believe they have obvious evidence that the effect is real. This is equally incorrect.

Stating that you cannot determine whether the effect is real or not because the sample isn't big enough is not stating either that the effect is real or not real. It's simply stating that you cannot draw any conclusions. Usually conclusions are being drawn when it is stated.

In fact, the original post is a good example. It was probably less than 50 innings and could simply be the product of 2-3 bad games.

[HRBaker]

Originally Posted by BPS

Yes, HOT is simply just the identification of past behavior (a series of good outcomes).

And, yes, outcomes in OOTP are determined by random variables. Yet, it is hypothetically the case that a series of past good outcomes can be partly due temporary boost in the probability of a good outcome.That is, some hotness could (hypothetically) be due to something other than pure randomness.

My guess is I'm wrong, but nothing the developers have said eliminates the above as a possibility.

Except Logic.

Player A (let's call him John Doe) has a card that represents a .300 BA and 30 HRs.

If John Doe were to get "hot" and bat .500 over 10 games, and there was a "trigger" in the game that upped his BA 10% until he had a 0-4 game... you've basically destroyed the game. John Doe's card no longer represents a .300 BA but maybe a .310 BA.

Every stat in the game HAS TO represent the players actual stats, or the whole concept goes down the drain.

[BPS]

Originally Posted by Orcin

Here's what I know: some of my cards perform better than the equivalent card on other teams, and some of my cards perform worse than the equivalent card on other teams. This variation is consistent and occurs over a large sample size. There hasn't been any credible explanation for this variation other than dumb luck.

Even if something is consistent with dumb luck, this is not evidence that dumb luck caused what has been observed.

But in any case, no one has shown that the results are, indeed, the result of dumb luck: no actual statistical evidence has been offered. What has been offered are general statements that "random stuff is going on and so what you see is the result of randomness. QED!" Not exactly a compelling argument, that.

[BPS]

Originally Posted by HRBaker

Except Logic.

Player A (let's call him John Doe) has a card that represents a .300 BA and 30 HRs.

If John Doe were to get "hot" and bat .500 over 10 games, and there was a "trigger" in the game that upped his BA 10% until he had a 0-4 game... you've basically destroyed the game. John Doe's card no longer represents a .300 BA but maybe a .310 BA.

Every stat in the game HAS TO represent the players actual stats, or the whole concept goes down the drain.

The mechanism I have suggested is not that past hotness increases future likelihood of success.

I have suggested that it is hypothetically possible that the "hotness" designation might the result of past temporary changes in the probability of a good outcome. And, a chance exists that this past temporary change might continue in the future (until some unspecified trigger occurs).

If the granting of hotness or coldness to a small number of player was determined randomly then this would not make the aggregate statistical output (that is, HRs, etc) not differ much from what it otherwise would have been. For every + impact you'd typically have a - impact.

In short, it would be easy to program in hotness and coldness into a baseball game so that no undesired impact would occur on the behavior of players.

[BPS]

Originally Posted by zrog2000

Stating that you cannot determine whether the effect is real or not because the sample isn't big enough is not stating either that the effect is real or not real. It's simply stating that you cannot draw any conclusions. Usually conclusions are being drawn when it is stated.

In fact, the original post is a good example. It was probably less than 50 innings and could simply be the product of 2-3 bad games.

I agree 100% with you. Invoking "small sample size" merely means we don't know. It does not mean the claimed effect isn't real. But, here, those swinging the cudgel of "small sample size" often seem to think they have smite the claims of their opponent (riding on the horse of a small sample).

If someone says, "we don't know" that is very different from "that is explained by random chance, you idiot."

[Orcin]

Originally Posted by BPS

Even if something is consistent with dumb luck, this is not evidence that dumb luck caused what has been observed.

But in any case, no one has shown that the results are, indeed, the result of dumb luck: no actual statistical evidence has been offered.

And no one has proven that that it is not dumb luck. So I will stick with my conclusion unless someone has evidence to the contrary.

[HRBaker]

Having a math background, I can't conceive of why a programmer would even want to instill a "trigger" of any sort. Randomness and the "law of averages" gives one all the possibilities needed to simulate reality.

[Matt Arnold]

Originally Posted by BPS

The mechanism I have suggested is not that past hotness increases future likelihood of success.

I have suggested that it is hypothetically possible that the "hotness" designation might the result of past temporary changes in the probability of a good outcome. And, a chance exists that this past temporary change might continue in the future (until some unspecified trigger occurs).

If the granting of hotness or coldness to a small number of player was determined randomly then this would not make the aggregate statistical output (that is, HRs, etc) not differ much from what it otherwise would have been. For every + impact you'd typically have a - impact.

In short, it would be easy to program in hotness and coldness into a baseball game so that no undesired impact would occur on the behavior of players.

Yeah, but we don't.

[BPS]

Originally Posted by Orcin

And no one has proven that that it is not dumb luck. So I will stick with my conclusion unless someone has evidence to the contrary.

I see: others are obliged to present evidence. You, however, feel comfortable making assertions without evidence.