Posted by crazystengel on 5/31/2011 7:43:00 PM (view original):
I actually prefer it if owners think certain players of theirs are "bad versions."  I'm in an OL whose regular season just finished.  The top 4 WW crapshooters in the league had win totals of 79, 70, 51, and 39.  Keep rollin' the dice -- you might come up with a player that Admin has "boosted" (instead of "poisoned")!

Posted by mixtroy on 5/31/2011 8:24:00 PM (view original):

Posted by crazystengel on 5/31/2011 7:43:00 PM (view original):
I actually prefer it if owners think certain players of theirs are "bad versions."  I'm in an OL whose regular season just finished.  The top 4 WW crapshooters in the league had win totals of 79, 70, 51, and 39.  Keep rollin' the dice -- you might come up with a player that Admin has "boosted" (instead of "poisoned")!

I guess the use of the WW was quite unfortunate for those guys, as I am sure they learned a valuble lesson. But not all WW users are created equal, as I often lead the league in WW transactions and still have success. You won't see me dumping three or four of my best, and most expensive, guys.  But  used prudently, the WW can definitely help a team.

Respectfully noted and understood Crazy.

Posted by contrarian23 on 5/31/2011 5:29:00 PM (view original):
This will likely not convince anyone, but here goes...

Microsoft Excel has a handy random number generator.  I just used it to simulate 100 "seasons" of 200 at bats, similar to the Jimmy Dykes example in the first post.  I modeled what a .300 hitter would do if he had a constant 30% chance of getting a hit in each at bat.

Across the 100 "seasons" he hit for anywhere between .175 on the low end and .365 on the high end.  By definition there are no "bad" versions here.  Every season, every at bat, is exactly the same.  Pure random variation, with no external causes whatsoever, can lead to very very dramatic differences in performance.

Of course, in WIS, there are many things that happen that I have not factored in: different pitching staffs, different parks, normalization, etc.  All of these will serve to increase the variation in expected performance.

My conclusion - and again I don't expect this to convince anyone else - is that a .175 batting average in 200 AB is not an unreasonable outcome.  It's not typical, but it's not at all outside the possibility of what pure random chance will produce.  It doesn't mean you got a bad draw, it doesn't mean that the player will continue to perform at that level.  You can replace him if you want; heck, I totally understand the emotion of it.  Assuming you replace him with a guy with similar "real" numbers, of course the new guy will do better than the old guy.  But the odds are very very likely that the old guy would ALSO have done better if you had left him in there.  My belief is there is no advantage to dumping a player for perceived underperformance, unless one of the drafting errors described by others in this post applies.

You can use this chart to see what to  expect a player to hit at a 95% confidence level.  For example if you expect a guy to hit .320 over 600AB, there is a 95% chance he will hit within 38 points of that expected average (.320 +/- 0.038 or .282 to .358). 

.

Posted by zubinsum on 6/1/2011 3:13:00 AM (view original):

Posted by contrarian23 on 5/31/2011 5:29:00 PM (view original):
This will likely not convince anyone, but here goes...

Microsoft Excel has a handy random number generator.  I just used it to simulate 100 "seasons" of 200 at bats, similar to the Jimmy Dykes example in the first post.  I modeled what a .300 hitter would do if he had a constant 30% chance of getting a hit in each at bat.

Across the 100 "seasons" he hit for anywhere between .175 on the low end and .365 on the high end.  By definition there are no "bad" versions here.  Every season, every at bat, is exactly the same.  Pure random variation, with no external causes whatsoever, can lead to very very dramatic differences in performance.

Of course, in WIS, there are many things that happen that I have not factored in: different pitching staffs, different parks, normalization, etc.  All of these will serve to increase the variation in expected performance.

My conclusion - and again I don't expect this to convince anyone else - is that a .175 batting average in 200 AB is not an unreasonable outcome.  It's not typical, but it's not at all outside the possibility of what pure random chance will produce.  It doesn't mean you got a bad draw, it doesn't mean that the player will continue to perform at that level.  You can replace him if you want; heck, I totally understand the emotion of it.  Assuming you replace him with a guy with similar "real" numbers, of course the new guy will do better than the old guy.  But the odds are very very likely that the old guy would ALSO have done better if you had left him in there.  My belief is there is no advantage to dumping a player for perceived underperformance, unless one of the drafting errors described by others in this post applies.

You can use this chart to see what to  expect a player to hit at a 95% confidence level.  For example if you expect a guy to hit .320 over 600AB, there is a 95% chance he will hit within 38 points of that expected average (.320 +/- 0.038 or .282 to .358). 

.