OOTP as not predictor, but PROPHET!! (poss. bug?)

[uabfan]

This happened a while back (still v13, though) but I'm in an online league and someone did a retrospective that brought it back to mind and, after another GM confirmed seeing this in a local league, I thought I would toss it to the user community.

The league is a fictional league and the "modern era" dates back to the 1995 season - we are in the middle of the 2008 season now. In '99, I drafted an OF who blazed through my minor leagues and had his rookie year in '02. In 2003 he produced the following stats:

157 G, 253 H, 46 2B, 50 HR, 137 R, 159 RBI, 7 SB, .413/.457/.749 (1.206 OPS)
141.0 VORP
+13.7 WAR

The bolded stats made him our league's first Triple Crown winner, his +13.7 WAR was a record until our last season, and his 141.0 VORP was greater than 13 of the 24 teams in our league that season. So I'm saying this was a statistical outlier for him (just his 2nd year) AND for our league. He singlehandedly carried my team to the playoffs.

When '04 Spring Training kicked off and OOTP produced it's Preseason Projections, I first noticed that my team predicted to drop 24 games from 2003 and back into 3rd in our division. Then I looked at the Predicted Top 10 Hitters and the player above was not included. At all. Like all my vets, I played him as a regular backup the first half of ST and then as a starter with more frequent breaks in the last half. On the next-to-last day of Spring Training, I get the news that he suffered a torn labrum diving for a ball in the outfield and his career was over. That was that.

The issue for the community is that I always thought it was a strange coincidence that the Preseason Predictions would exclude a hitter coming off the single greatest offensive season in the history of our league and then he just happens to have a career-ending injury. When I brought it back up recently, another GM mentioned he had a local league where his top 3 starters were among the best in the league, candidates for Cy Young, etc. OOTP Preseason Predictions come out one year and the Predicted Top 10 Pitchers has his #1 and #3, but curiously excluded his #2 starter. Sure enough, his first start of the season he suffers a career-ending injury.

Now, I don't think it's a game-breaker that OOTP would know before a season started that a player was going to suffer a severe or career-ending injury (unless, of course, AI GMs know that, too, before the season starts). I would even think the engine MUST know before the fact that Joe Nobody has a 92% likelihood of suffering a career-ending injury from data hidden from users (probability tables or what have you). But for the engine to take that into account when making Preseason Predictions can't be the way that feature was designed. If so, anyone with a stud hitter or pitcher conspicuously absent when the Predictions come out could try to swing a quick trade to someone not paying attention or even make a move for better quality backup. Basically, if this is indeed figured into Predictions, that's free, unrealistic Nostradamus info for GMs. Anyone else seen this?

[Honorable_Pawn]

This is disturbing. If this is true and since you posted it maybe you should make sure that it goes up.

[The Wolf]

File a bug report on this, please.

[Jordan]

i would hope this is a coincidence. and not a design.

[uabfan]

I don't seem to have appropriate rights to post on the "Bug Reports" thread. Is there an approval step I need to go through?

[uabfan]

Originally Posted by Jordan

i would hope this is a coincidence. and not a design.

Helluva coincidence. Like I said in the OP, it wouldn't be surprising to learn that the OOTP engine has access to injury probabilities for every player before a season starts. But since that kind of information is not accessible to players, I'm not sure it should be considered in Preseason Predictions.

Now, if this guy hadn't been coming off that historic season all I would have seen is my team predicted to tank in terms of wins/losses (which we definitely did). It just happened to be that he was curiously absent from the Predicted Top 10 Hitters list and then suffers a career-ending injury before Opening Day. Definitely made me go "Hmmm..."

[Charlie Hough]

What was the prediction for him going into the previous season? And what were his ratings going into the season before he suffered the injury? I'm willing to bet that the predictions are based on ratings and not stats.

If you have AI evaluation turned on, and you have included stats, I suppose they could count for something, but that's for the AI. It doesn't mean that OOTP is going to view it that way for prediction purposes.

My guess is that his stats are not considered at all in the predictions and his ratings were compared to the other hitters in the league. If his ratings are borderline to be in the top 10, then I think this is merely a coincidence.

[uabfan]

Originally Posted by Charlie Hough

What was the prediction for him going into the previous season? And what were his ratings going into the season before he suffered the injury? I'm willing to bet that the predictions are based on ratings and not stats.

If you have AI evaluation turned on, and you have included stats, I suppose they could count for something, but that's for the AI. It doesn't mean that OOTP is going to view it that way for prediction purposes.

My guess is that his stats are not considered at all in the predictions and his ratings were compared to the other hitters in the league. If his ratings are borderline to be in the top 10, then I think this is merely a coincidence.

I can't remember if he was projected to be amongst the Top 10 Hitters before 2003 (his monster season at age 26). His ratings were:
1/1/2003: vR: 20-15-19-10-20; vL: 18-13-16-8-18

Then, when he suffered his career-ending injury on 3/24/04 his ratings were (age 27):
3/24/2004: vR: 20-15-20-12-20; vL: 18-13-18-10-18

(Keep in mind that our league settings cut ratings off at 20...I could get the Commish to look at his raw ratings if they're still available)

His 2002 stats (age 25) were:

131 G, .281/.335/.484, 21 HR, 66 RBI, 70 R, 5 SB, 19.6 VORP, +1.0 WAR

I do see that my team was predicted to go 92-70 in 2003 (which we did) and then 69-93 in 2004 (which we did).

[Tram2Whitaker]

It seems like quite a bit of work, but to research this you would have to save the league at the point in time that you noticed that the predictions no longer include him. Then, simulate the same season, multiple times. If the outcome is the same, then yes, there is an unseen stat (probably correlates to injury history) If the end result is the same, but the means by which it comes to that end differs (such as getting injured in May rather than ST), then there probably is a stat that can be quantified. If the result cannot be consistently duplicated, then it's probably just coincidence.

Not like you can do that now, nor would it be easy to catch when it might happen again; but I would do this before determining whether it was a bug or not.

[RchW]

Originally Posted by uabfan

I do see that my team was predicted to go 92-70 in 2003 (which we did) and then 69-93 in 2004 (which we did).

This is strange. I find the predictions variable at best. In my latest season KC finished 95-67 and was predicted to be 67-95. It got 2 of 3 NL division winners but none of the AL. Even where it was close to the right standings the games won could be off as much as 16.

[uabfan]

Originally Posted by RchW

This is strange. I find the predictions variable at best. In my latest season KC finished 95-67 and was predicted to be 67-95. It got 2 of 3 NL division winners but none of the AL. Even where it was close to the right standings the games won could be off as much as 16.

It is strange...unless you consider, hypothetically, that the predictions were generated taking into account that my team would be without a player that generated 141.0 of my team's total 209.9 VORP the previous year.

[RchW]

Originally Posted by uabfan

It is strange...unless you consider, hypothetically, that the predictions were generated taking into account that my team would be without a player that generated 141.0 of my team's total 209.9 VORP the previous year.

I'd have to see way more evidence before I buy that.

[uabfan]

Originally Posted by RchW

I'd have to see way more evidence before I buy that.

Well, that was kind of my point in bringing it up on this forum. Like I said, it was a while back so I obviously didn't think it was that urgent (I was too busy flipping tables over the injury itself, tbh). Plus, since it only lists the top 10 hitters and top 10 pitchers for each sub-league, it's going to be very rare that the returning league MVP suffers a career-ending injury in ST at the age of 26. I just remember being shocked that we were predicted to fall off so dramatically AND that this player was absent from the top 10 hitters...then, after the injury, I made the connection and it seemed very curious.

I can say that I've never had a team other than that one be predicted to nosedive like that...


Sent from my iPad using Tapatalk HD

[RchW]

Originally Posted by uabfan

I just remember being shocked that we were predicted to fall off so dramatically AND that this player was absent from the top 10 hitters...then, after the injury, I made the connection and it seemed very curious.

I can say that I've never had a team other than that one be predicted to nosedive like that...


Sent from my iPad using Tapatalk HD

I agree. I'll go back over some past seasons to see if that occurred in my league.

I've checked several seasons and found no players who could be described as good/great who did not appear on the pre-season list who subsequently got injured. I'm coming to the preliminary conclusion that the pre-season list is based on the previous season. Not appearing on it is a better predictor of good batting results than being on it.