Posted by mitchrapp on 6/26/2010 4:46:00 PM (view original):
It's not all about Math. If it was then the .250 hitter year after year you'd keep playing and paying because the same exact player with the same ratings is batting .300 year after year.but eventually you'll drop that player or bench or trade him. It makes sense.

Or more importantly why does a career .300 hitter, hit .222 one year with the same ratings and park. Saying Math is saying they will allays be consistent. And that is not always the case, we've all seen it. I get what you want to believe but it's not all about anything. It's about many different things.

Yeah, that's still math.

Posted by mitchrapp on 6/26/2010 4:46:00 PM (view original):
It's not all about Math. If it was then the .250 hitter year after year you'd keep playing and paying because the same exact player with the same ratings is batting .300 year after year.but eventually you'll drop that player or bench or trade him. It makes sense.

Or more importantly why does a career .300 hitter, hit .222 one year with the same ratings and park. Saying Math is saying they will allays be consistent. And that is not always the case, we've all seen it. I get what you want to believe but it's not all about anything. It's about many different things.

You don't know what math is if you believe any of that nonsense.

It's more about opportunity.   First, if you even test the line-up link I posted earlier, and you believe they know how to do math, you'll see that line-up doesn't make that much difference.  Second, regardless of ratings, most of us want the player who's hitting .300 to get one more AB than the guy hitting .210(and yes, I'm just using one stat for simplicity).    Third, line-ups aren't etched in stone.    You can change them from game to game.

So, with all this in mind, you can move the .300 hitter up and the .210 down until the trend starts to reverse.   Or you can leave them as is because ratings drive the game.  It's just a personal preference.   I know some of you love the idea of me batting my 88 stud seventh because he's hitting .210.  I love the idea of you leaving your 88 stud hitting .210 in the three hole. 

To put it another way, if I have the opportunity to adjust my rotation to pitch lefties or righties against you after the A/S break, I look at your stats vs. LH/RH.   I don't look at your player's ratings.   Ratings drive the game but the results determine the winner.

Posted by MikeT23 on 6/26/2010 5:40:00 PM (view original):
To put it another way, if I have the opportunity to adjust my rotation to pitch lefties or righties against you after the A/S break, I look at your stats vs. LH/RH.   I don't look at your player's ratings.   Ratings drive the game but the results determine the winner.

Agree that any and all maneuvering we do with lineups doesn't likely have a ton of impact.

That said, you (perhaps inadvertently) struck the nail on the head with your last statement.

"Ratings drive the game but results determine the winner."

You seem to recognize that the prior game's results do not directly impact success or failure in the following game (unlike those who believe in programmed streaks), so it doesn't make any logical sense to make lineup decisions based on them. It is ratings that are directly predictive of future performance, far more so than statistics (even if you could control for all of the variables involved, which you cannot).

Just to be perfectly clear, I'm not talking about the situation where you bench the 80/80/80/80/40 guy hitting .250 in favor of the 70/70/70/70/70 guy hitting .300. That's a case where perhaps the value of eye is underrated initially, or the value of splits overrated. There's a judgment call in connection with that where a possible error is made.

Benching the 80/80/80/80/80 guy hitting .250 in favor of the 70/80/80/80/80 guy hitting .300, with every other rating being identical, is simply not a recipe to maximize success in the long term.

It takes a lot for me to bench a stud.   Not so much to move him in the line-up. 
http://www.whatifsports.com/HBD/Pages/Popups/PlayerProfile.aspx?pid=1469204 has hit 2nd for me for the last two seasons.  He's hitting second for me against lefties this season.  But, vs RH, he's .225/.298/.346.   It makes no sense to keep him hitting 2nd.    If those numbers get better, I'll move him back up.  If not, I'm more than happy to have someone else fill that slot in the order.

Posted by zbrent716 on 6/26/2010 5:53:00 PM (view original):

Posted by MikeT23 on 6/26/2010 5:40:00 PM (view original):
To put it another way, if I have the opportunity to adjust my rotation to pitch lefties or righties against you after the A/S break, I look at your stats vs. LH/RH.   I don't look at your player's ratings.   Ratings drive the game but the results determine the winner.

Agree that any and all maneuvering we do with lineups doesn't likely have a ton of impact.

That said, you (perhaps inadvertently) struck the nail on the head with your last statement.

"Ratings drive the game but results determine the winner."

You seem to recognize that the prior game's results do not directly impact success or failure in the following game (unlike those who believe in programmed streaks), so it doesn't make any logical sense to make lineup decisions based on them. It is ratings that are directly predictive of future performance, far more so than statistics (even if you could control for all of the variables involved, which you cannot).

Just to be perfectly clear, I'm not talking about the situation where you bench the 80/80/80/80/40 guy hitting .250 in favor of the 70/70/70/70/70 guy hitting .300. That's a case where perhaps the value of eye is underrated initially, or the value of splits overrated. There's a judgment call in connection with that where a possible error is made.

Benching the 80/80/80/80/80 guy hitting .250 in favor of the 70/80/80/80/80 guy hitting .300, with every other rating being identical, is simply not a recipe to maximize success in the long term.

+1

The last 2 paragraphs are key here.  We are not talking about misinterpreting ratings (zbrent's first example).  We're talking about hitters who are better in every way, shape and form, i.e. Zbrent's second example.

Posted by MikeT23 on 6/26/2010 5:59:00 PM (view original):
It takes a lot for me to bench a stud.   Not so much to move him in the line-up. 
http://www.whatifsports.com/HBD/Pages/Popups/PlayerProfile.aspx?pid=1469204 has hit 2nd for me for the last two seasons.  He's hitting second for me against lefties this season.  But, vs RH, he's .225/.298/.346.   It makes no sense to keep him hitting 2nd.    If those numbers get better, I'll move him back up.  If not, I'm more than happy to have someone else fill that slot in the order.

Next game vs RH, your stud's ratings make him no less worthy of batting 2nd than they did the game before. If you are replacing him with someone with *different* ratings (better in one thing or another) because you have determined that you were over-valuing one aspect of the stud's ratings, that's one thing. If you are hitting someone with lesser ratings 2nd based on stats, you're falling prey to an aspect of the Gambler's Fallacy, just as much as the person who defends keeping the stud 2nd because he is "due" to hit .400 the rest of the season to even things out. You're both equally wrong.

The only aspect of a player that directly impacts success or failure in HBD is his ratings. Accepting that and still basing decisions on anything else is willful ignorance.

No, we're really talking about refusing to make changes in order to better your team.   There are several things we agree upon.  Players over/underperform from season to season.  Ratings drive the game.   A single season is a small sample size.  Line-ups do not make that big of a difference.   For me, I'll adjust my line-up based on results because, as we know, 1) a season is a small sample size 2) players underperform for no apparent reason.  You, on the other hand, are willing to wait an entire season to find out if a player is going to underperform for the entire season.    IMO, that makes no sense.  In your opinion, that's the perfect way to run your team.   Unless you disagree with one of the bolded, italicized statements, there really isn't a reason to argue.

Here is Haywood Twitchell. 

Haywood is a lefty and has almost identical ratings versus lefties and righties.  82 vsL and 80 vsR.  Keep in mind he is left handed so he is at a platoon disadvantage against 95% of left handed pitchers (because their vsR ratings is higher than their vsL).

We are in the playoffs this year and here are Haywood's full season splits versus lefties and righties (these numbers are in pitchers park Burlington)

Versus Lefties - 170 AB - .335, .389, .653
Versus Righties - 351 AB - .288, .351, .487

So is he slumping or not?  Or just slumping versus righties?  Is he hot versus lefties?  Why would he be slumping versus only righties?  So on top of the streaks you guys think are built into this game do you also think players can slump versus only right handed pitchers?  His numbers versus lefties are far ABOVE what you would expect out of him in Burlington and his numbers versus righties are a good deal below what you'd expect.

I can't think of a better example of randomness than this. 

Posted by zbrent716 on 6/26/2010 6:03:00 PM (view original):

Posted by MikeT23 on 6/26/2010 5:59:00 PM (view original):
It takes a lot for me to bench a stud.   Not so much to move him in the line-up. 
http://www.whatifsports.com/HBD/Pages/Popups/PlayerProfile.aspx?pid=1469204 has hit 2nd for me for the last two seasons.  He's hitting second for me against lefties this season.  But, vs RH, he's .225/.298/.346.   It makes no sense to keep him hitting 2nd.    If those numbers get better, I'll move him back up.  If not, I'm more than happy to have someone else fill that slot in the order.

Next game vs RH, your stud's ratings make him no less worthy of batting 2nd than they did the game before. If you are replacing him with someone with *different* ratings (better in one thing or another) because you have determined that you were over-valuing one aspect of the stud's ratings, that's one thing. If you are hitting someone with lesser ratings 2nd based on stats, you're falling prey to an aspect of the Gambler's Fallacy, just as much as the person who defends keeping the stud 2nd because he is "due" to hit .400 the rest of the season to even things out. You're both equally wrong.

The only aspect of a player that directly impacts success or failure in HBD is his ratings. Accepting that and still basing decisions on anything else is willful ignorance.

There comes a time in every player's season when his ratings don't determine where he bats in the line-up.   His stats do.  As I said, I'm not benchng a guy I believe is better than his replacement because he's not producing.  But I'm not hitting him at the top of the line-up because his ratings say "He's my two hitter."    I believe, at some point, you have to get results.   

Posted by MikeT23 on 6/26/2010 6:07:00 PM (view original):
No, we're really talking about refusing to make changes in order to better your team.   There are several things we agree upon.  Players over/underperform from season to season.  Ratings drive the game.   A single season is a small sample size.  Line-ups do not make that big of a difference.   For me, I'll adjust my line-up based on results because, as we know, 1) a season is a small sample size 2) players underperform for no apparent reason.  You, on the other hand, are willing to wait an entire season to find out if a player is going to underperform for the entire season.    IMO, that makes no sense.  In your opinion, that's the perfect way to run your team.   Unless you disagree with one of the bolded, italicized statements, there really isn't a reason to argue.

1) Absolutely.

2) Players underperform for no apparent reason. Apparent reason. I believe that there IS always a reason, even if it not apparent. Maybe the stud has had more hits stolen from him by great plays by the opposing team. Maybe the stud has been matched up against the best RP from the other team more often. Maybe it's something else.

But there IS some reason and unless you are predicting that same (unobserved) reason to exist moving forward, ratings still drive the failure/success in the future.


Possibly related, but in writing this and trying to make it as simple as possible, I did think of something interesting that I hadn't considered before. Note that I do not believe the engine works this way - in fact I am nearly certain it doesn't - but I am not 100% certain and it is a possibility. The specialists (with their far from perfect usage) are generally, I think, called in to face tougher batters.

That is, a RH specialist isn't going to be called in to face the #9 RH batter who is a defensive SS specialist with no bat. He may, however, come in to face the RH stud batter (stud by ratings, regardless of stats). Now, the specialists don't log nearly enough PA/IP to impact anything significantly, but what about Setup A vs B? Could it be possible, in theory, that the SimAI would bring in the Setup A if you left the under-performing stud in (because it looks only at ratings) but might bring in the Setup B if the batter was crappy (based again on ratings, not stats). So, over the long run, the backup/scrub consistently faces the lesser pitcher, so has more success.

I'm pretty sure it doesn't work this way, and that Setup A vs B is dictated by score, inning, etc. rather than match-up, but it was an interesting thought that I hadn't considered before.

Posted by MikeT23 on 6/26/2010 6:07:00 PM (view original):
No, we're really talking about refusing to make changes in order to better your team.   There are several things we agree upon.  Players over/underperform from season to season.  Ratings drive the game.   A single season is a small sample size.  Line-ups do not make that big of a difference.   For me, I'll adjust my line-up based on results because, as we know, 1) a season is a small sample size 2) players underperform for no apparent reason.  You, on the other hand, are willing to wait an entire season to find out if a player is going to underperform for the entire season.    IMO, that makes no sense.  In your opinion, that's the perfect way to run your team.   Unless you disagree with one of the bolded, italicized statements, there really isn't a reason to argue.

Well actually we're talking about making changes in order to make your team worse.

And I disagree with one of the bolded, italicized statements.

Players over/underperform from season to season can be more precisely called "players over/underperform from atbat to atbat."

At bats are independent events from one another, like rolls of the dice.  Sure your guy 88 rated stud is hitting 1 for his last 25 but, and here is where we differ, this offers no predictive ability for his next at bat.  Only the ratings predict the next at bat, if you can read the ratings right.

You don't know what a player will do in the future based on his "slump".  When does the slump end, how can you tell?  What if he comes out of the slump and you finally move him back to the 3 slot and he goes into a slump again, when can you tell a player is back in a slump?.  Slumps are more noticeable at the beginning of seasons because we notice them in the stats easier.  Going into the playoffs do you go through game logs and plot what your players have done in the last 20 or 40 games and adjust accordingly?

Posted by antonsirius on 6/26/2010 1:20:00 PM (view original):

Posted by schuyler101 on 6/26/2010 12:30:00 PM (view original):
And oddly enough it is the only right answer

Did you seriously just tell people the Gambler's Fallacy isn't a fallacy?

Haha, yeah whoops.  Didn't read through it.

That is a pretty clear example of the gambler's fallacy

I read as far as "Leave him batting 3rd because the "law of averages" says..."

I just assumed that one stated my position...that the law of averages says he'll hit comparable to his baseline going forward.