Arm/range studies

Postby The Biomechanical Man » Wed Jun 11, 2008 7:08 pm

gkhd,
Thanks for sharing your insight. Computing the effects on Wins for players with good or bad fielding is strongly related to TSN [i:ac4d524ca6]pricing[/i:ac4d524ca6] of these players. TSN [i:ac4d524ca6]might[/i:ac4d524ca6] :roll: be placing more or less player value with fielding in different online card sets.

So, the finding that better arms lead to more wins and titles might be true only with the pricing set of the card set (ATG1, ATG2, ATG3, 2007, etc.) you played with.
The Biomechanical Man
 
Posts: 55
Joined: Tue Jul 03, 2012 2:34 pm

STATISTICS 101

Postby The Biomechanical Man » Thu Jun 12, 2008 6:16 pm

Dean emailed me his data, and I run a statistical analysis. (Specifically, I ran an ANOVA for each variable - cf arm, rf arm, etc. - and then post-hoc tukey t-tests. I used p<0.05 for significant.) Here's what we have:


[color=darkred:cbdbca0737][b:cbdbca0737][u:cbdbca0737]STATISTICALLY SIGNIFICANT DIFFERENCES[/u:cbdbca0737]

C range 1: -13 runs
C range 3 (original): 0
C range 5: +7 runs

C arm -2: -10 runs
C arm +1 (original): 0
C arm +4: +8 runs

CF arm -3 (original): 0
CF arm 0: +2 runs
CF arm +3: +17 runs[/b:cbdbca0737] [/color:cbdbca0737]


[u:cbdbca0737]STATISTICALLY NON-SIGNIFICANT[/u:cbdbca0737]

LF arm -5: -6 runs
LF arm -2 (original): 0
LF arm +1: 6 runs

RF arm -3: -2 runs
RF arm 0 (original): 0
RF arm +3: +5 runs


The post-hoc tests for catcher and centerfield data showed significant differences only between the two extremes of the three conditions.


What does this gibberish mean? :P Well, first let me point out that there is [i:cbdbca0737]statistical[/i:cbdbca0737] significance and there is [i:cbdbca0737]practical[/i:cbdbca0737] signficance. Statistical significance is how sure we are that the differences found really exist, not just dumb luck. Practical difference is whether the difference is important to you.

The statistics says that we are more than 95% certain that changes in catcher arm, catcher range, and centerfield arm do affect how many runs your team will allow. Dean's data was unable to prove that LF and RF arm affect how many runs a team allows (these arms might truly have no effect on runs allowed, or perhaps Dean just didn't run enough sims.)

Practical significance is in the eye of the beholder. Is it important if a change in defense leads to 10 more (or less) runs allowed during a season? That's up to each manager.
The Biomechanical Man
 
Posts: 55
Joined: Tue Jul 03, 2012 2:34 pm

Postby Mean Dean » Fri Jun 13, 2008 12:36 pm

Glenn, thanks for doing the work on that; much appreciated. I'm not a statistician, so underlying my work is the "common sense" assumption that changing anything at all about the card has to have at least some miniscule effect somewhere ;) But you're right, strictly speaking, the initial assumption should be that nothing has any effect, unless and until you prove that it does.

Re: translating range ratings into runs: I think/hope that my "offense vs. defense" article gets at this issue just as well as running hundreds of simulated seasons would, and in a way that is a lot less tedious for me to calculate ;)
Mean Dean
 
Posts: 55
Joined: Tue Jul 03, 2012 2:34 pm

Postby SteadyEddie33 » Sun Jun 15, 2008 3:24 pm

I'm trying to catch up on the analytical discussions, so I'm very late, lol. However, it seems as though this was a very important post/point that was overlooked. In leagues where guys are constantly tweaking/micromanaging, such as matchup SPs and the like, you'd better believe that seeing a couple "+" arms in the outfield has an effect on the strategy being set as "aggressive or conservative" for base advancement.
More opposing managers will choose to test the "+" outfield, creating the negative feedback/exponential growth effect. Remember, we don't get to choose to run more on the CF or LF, it's a decision that holds for the entire unit, based on the individual ratings.

So the question seems to become, psychologically, does having a -4 in RF effectively balance a +1 CF and/or LF??

[quote:0c1a9adf02="Bbrool"]Dean,

You will find that the catchers arms impact the game very similarly to the outfielder arms. I went back and reviewed 30 to 40 plus seasons for each team in the league, and the reason this holds true is that a good catcher can only get finitely better than the average catcher (i.e. if an average catcher (-1) arm, allows a mix of SB/CS that estimates to say 15 runs a season, then the best catcher can only be 15 runs better than your baseline (in fact regardless of the arm in the -4 to -1 range, the sb% is roughly the same, it is the attempts that go down).

However, on the other end of the spectrum, as the arm gets worse, more and more runners steal, and in addition the SB % increases so the combination creates an exponential growth in runs allowed for the worse catchers.

I would assume that this analogy holds true for OF arms as well.

Bbrool[/quote:0c1a9adf02]
SteadyEddie33
 
Posts: 55
Joined: Tue Jul 03, 2012 2:34 pm

Postby Mean Dean » Mon Jun 16, 2008 10:51 am

I don't see how that makes sense. It would only make sense if the aggressiveness setting means "go for the extra base X percent of the time, regardless of your chance to make it." But I don't think there's any chance it means that. Setting it higher has to either mean A) that you go more often given a certain success chance (eg, conservative means you'll go 25% of the time on a 60% chance, aggressive means you'll go 50% of the time on a 60% chance), or B) that you'll go on lower chances (eg, conservative means you'll try to go 1st to 3rd with one out on a 75% chance or better, aggressive means you'll try to go 1st to 3rd with one out on a 50% chance or better.) Those are technically a little different, but they amount to the same thing: there would be no reason to change the setting against bad arms. What you want against bad arms is not to run more often given the same percentage (that'd be case A), and [i:d7ab9d8ca4]certainly[/i:d7ab9d8ca4] not to run at lower percentages (that'd be case B). What bad arms do is give you more opportunities to run at the percentages where it's profitable to run. That takes care of itself, if you just leave the setting alone. (I believe case B is how it actually works, BTW. I really need to stop procrastinating on the managerial settings studies that will hopefully settle the matter.)
Mean Dean
 
Posts: 55
Joined: Tue Jul 03, 2012 2:34 pm

Previous

Return to Strategy

Who is online

Users browsing this forum: No registered users and 2 guests