Recalc Question

[David Watts]

Random Debut league using 3 year recalc. The option for double current years stats is unchecked.

Nomar Garciaparra hits .325 the first season of my league. Follows that up with a season in which he hits .334. Along comes the next season and Nomar is sitting on his teams bench as a backup to Wayne Causey.

Now, the season in play would be one in which Nomar only played in 21 games in real life, batting 91 times. He hit .289. But, that year was sandwiched between a season in which he hit .372 and one in which he hit .310. Nomar played in 140 games the .372 season and a 156 the .310 season. So his recalc 3 year average should be something like .340.

It appears to me that the computer is factoring in Nomar's real life injury season when it comes to making its lineup This is exactly what I was trying to avoid when I decided to use 3 year recalc. Could it be due to me having selected the option to use real stats? I want to soften the effects of real life injuries on my replay so guys don't go from Cy Young winner to bottom of the bullpen due to poor seasons in real life due to injuries.

I really thought using 3 year recalc would be the answer.

[Charlie Hough]

Are you using the option to use real life lineups and transactions? Maybe that is why Nomar is being kept out.

Or this may be a problem with your AI evaluation settings.

What are Nomar's ratings? And how much weight are you placing on stats vs. ratings in AI player evaluations?

But the stats are based on the simulated stats of the current season along with the previous two seasons, depending on your settings. So the real life stats shouldn't be driving this, and the AI can never base a lineup decision based on real life usage unless you're using real life lineups and transactions.

In the end, sometimes the AI makes some dumb decisions on player personnel. But I would check your settings first.

[David Watts]

Originally Posted by Charlie Hough

Are you using the option to use real life lineups and transactions? Maybe that is why Nomar is being kept out.

Or this may be a problem with your AI evaluation settings.

What are Nomar's ratings? And how much weight are you placing on stats vs. ratings in AI player evaluations?

But the stats are based on the simulated stats of the current season along with the previous two seasons, depending on your settings. So the real life stats shouldn't be driving this, and the AI can never base a lineup decision based on real life usage unless you're using real life lineups and transactions.

In the end, sometimes the AI makes some dumb decisions on player personnel. But I would check your settings first.

It will change a players role if their real life stats don't add up. For example, in my last random debut, Chris Carpenter won back to back Cy Young awards. The following season he was at the bottom of his teams bullpen as a mop up guy. He remained there for 2 seasons. These were 2 seasons in which he was injured in real life. I was using 1 year recalc for that league. That's why I thought 3 year recalc would be the better option.

Random debut league so real lineups don't exist.

[Charlie Hough]

With three-year recalc, there is no way that Nomar's single limited season is causing the lineups to change simply because he played only 21 games in real life.

If his recalc ratings dropped off dramatically for some reason because that season was absolutely terrible, then it's possible that the AI has chosen to use someone else.

But the ratings dropoff would have to be significant, and this is very unlikely for two reasons: one is that one season out of three is usually not enough to cause a massive change. Also, if you're using the setting to adjust ratings for fewer than X number of at bats and your setting is for at least 100 or fewer at bats, then his ratings should not suffer much anyway, even if he had a poor season in his 91 at-bats.

My guess is that his ratings must be similar to or worse than Causey's, and your AI evaluation settings are leading to the AI choosing Causey because most of the weight is on ratings and current season stats.

[David Watts]

Originally Posted by Charlie Hough

With three-year recalc, there is no way that Nomar's single limited season is causing the lineups to change simply because he played only 21 games in real life.

If his recalc ratings dropped off dramatically for some reason because that season was absolutely terrible, then it's possible that the AI has chosen to use someone else.

But the ratings dropoff would have to be significant, and this is very unlikely for two reasons: one is that one season out of three is usually not enough to cause a massive change. Also, if you're using the setting to adjust ratings for fewer than X number of at bats and your setting is for at least 100 or fewer at bats, then his ratings should not suffer much anyway, even if he had a poor season in his 91 at-bats.

Oh, I'm using the default settings for make hitters bad.

My guess is that his ratings must be similar to or worse than Causey's, and your AI evaluation settings are leading to the AI choosing Causey because most of the weight is on ratings and current season stats.

Nomar is rated higher than Causey across the board. Even rated higher defensively. Nomar is actually listed as the starting SS in the teams lineup against leftys, but even then the game will often start Willie Green over him. Strange. It will be interesting to see if Nomar is back to starting next season.

[David Watts]

Originally Posted by Charlie Hough

With three-year recalc, there is no way that Nomar's single limited season is causing the lineups to change simply because he played only 21 games in real life.

If his recalc ratings dropped off dramatically for some reason because that season was absolutely terrible, then it's possible that the AI has chosen to use someone else.

But the ratings dropoff would have to be significant, and this is very unlikely for two reasons: one is that one season out of three is usually not enough to cause a massive change. Also, if you're using the setting to adjust ratings for fewer than X number of at bats and your setting is for at least 100 or fewer at bats, then his ratings should not suffer much anyway, even if he had a poor season in his 91 at-bats.

My guess is that his ratings must be similar to or worse than Causey's, and your AI evaluation settings are leading to the AI choosing Causey because most of the weight is on ratings and current season stats.

What makes things even more odd is, I'm using neutralized stats. So, Nomar's recalc at bat totals for the injury season are actually boosted to I want to say 250. This is just one of those occasions where I'm left scratching my head. It's early in the season, so I will watch to see if somewhere along the line Nomar bumps Causey out of his starting role.

I am using lefty/righty splits and I know when it comes to historical they are completely fictional. I wonder if this could have something to do with it?

If you have any suggestions I would love to hear them. I know real life injuries and decline when using recalc can lead to players going from stars to duds in the off season, but I really thought 3 year recalc would soften this some.

[actionjackson]

Edit: never mind. Memo to self: next time read all posts before trying to help out.

[Charlie Hough]

Hmm. Well, I'm left with only a few possibilities that I can imagine here.

One is that the manager's strategy preferences are driving this. Does the AI manager favor speed or defense over other attributes? Compare the players involved with the manager's strategy preferences and see what you find.

Another is that the AI is seeing some hidden ratings or a total ratings picture that is beyond what you're considering. So, for example, it may calculate a higher overall value for these other players ahead of Nomar because their calculated total values are higher than his.

For example, when evaluating trades, the AI uses this overall number, so maybe it's doing the same thing here and combining defense, position, batting ratings, and hidden ratings to arrive at its own conclusions. But, since Nomar supposedly has higher ratings across the board, I suspect that this won't be the explanation. But maybe he has some ratings in a certain category that the AI doesn't like for strategy purposes.

As you mentioned, maybe the righty/lefty splits and your lineup settings are affecting things. Are you using traditional lineup AI or sabermetric AI, and what are Nomar's ratings vs. lefties and righties compared to these other guys?

Beyond this, it may simply be a case of the AI making dumb decisions, which I have seen before, but this can often be mitigated by the right settings.

If all else fails, you could use 7-day lineups, depth charts, and settings that disallow AI lineup changes. Or there is an option to force a player to start at a certain position.

However, I've seen other threads where people say that the AI sometimes ignores or works around this, but that was with past versions of OOTP. I would assume that these options work properly these days.

[David Watts]

Charlie, I'm thinking the AI might know its stuff. Garciaparra started to get more playing time, but wow is he stinking. He's hitting .157 for the season. Causey isn't doing much better. He's hitting .193.

[Charlie Hough]

Well, it depends on the sample size. How many games and at-bats are we talking about? It will probably shake out in the end, though sometimes there are exceptions.

In my fictional league that I am running, one of my players was one of the best all-around hitters in the league in year one, and he established league records. He was injured and missed a sizable portion of the second season, and he saw a dropoff in average and slugging despite having no real drop in ratings.

However, in his third season, he has inexplicably hit for a much lower average than year two and certainly compared to year one. His slugging is around the same as year two, but he has done quite poorly with runners on base and has grounded into a vastly higher rate of double plays, all despite having no dips in ratings.

This could be a topic for a separate thread, but his WAR and OPS have plummeted. This is over the course of 100+ games now. It's absolutely puzzling especially because the game has never designated him as being on a cold streak this season, and he plays consistently well below his expected output.

He went from being a .338 hitter with an OPS of 1.019 and WAR of 6.4 in year one to a .260 hitter with 0.815 OPS and 1.6 WAR in year three.

I just don't see how this could be happening because there are no appreciable ratings drops. He's gotten hurt a couple of times, so maybe that has had some invisible effect on him. But it really aggravates me because there is nothing in the game's ratings to indicate that he should be suffering from this kind of drop.