Advanced Analytics - Defense Topic

dw172300
Posts: 1173 (0)
Hall of Famer
1 182 5 92
Rookie
I've had some luck using advanced analytics to help me gauge batting statistics. Basically using a couple of formulas that use Runs Created along with a few other metrics to arrive at VORP and WAR. Has anyone had any luck being able to quantify the value of their players defensive stats versus a league average or replacement level player at each position?

I've seen some people who talk about using POs and Assists to do so, but shy away from those because those numbers are somewhat dependent on how many K's, Groundballs, and Flyballs each of your pitcher tends to throw. Has anyone used RF and F% to put together any basic analyses? I'd be interested in your tips on how to get started if so. Thanks.
6/4/2012 6:50 PM
dyoungquist
Posts: 866 (0)
Hall of Famer
180
Rookie

Range Factor (RF) is dependent on how many groundballs vs flyballs your pitchers give up and is a totally useless stat in my opinion.  You can have a SS with an above average range rating but if his pitchers give up a lot of flyballs he might not even make the top 5 in RF at SS.

6/4/2012 8:10 PM
deanod
Posts: 3152 (0)
Hall of Famer
115 18 1 119 1 52
Prospect
stick to +/- plays and errors bro.  everything else is meaningless
6/4/2012 8:16 PM
dahsdebater
Posts: 10046 (3)
Hall of Famer
284 293 18
Pro
I don't think there's a great way in this game to accurately quantify range.  I think the best you could do would be to use +/- plays as a proportion of total chances.
6/4/2012 11:57 PM
rangerup
Posts: 682 (0)
All Star
2 42
Prospect
Look at range rating.  That quantifies it.
6/5/2012 12:46 AM
dw172300
Posts: 1173 (0)
Hall of Famer
1 182 5 92
Rookie
I think you guys are missing my point. I'm looking to put a value to the defensive stats beyond those metrics themselves. Has anyone used the defensive stats available in the game to calculate VORP or WAR? I've done it with offensive numbers but not the defensive ones.
6/5/2012 1:01 AM
kahrtmen
Posts: 307 (0)
Hall of Famer
6 1 78
Prospect
Like deanod says, you can do it with +/- and errors.  Each extra play made (or error avoided) saves about .65-.75 runs.
6/5/2012 1:55 AM
MikeT23
Posts: 58012 (0)
Hall of Famer
87 1 47 1 189
Prospect
I use a very simply formula using PO, A, +, - and E for OF and A, +, - and E for IF then divide by Innings played to determine how many "positive" plays a player makes per inning.   It's not advanced by any means but it helped me realize that a heavy-hitting rock won't do much damage in RF.   And that you better be strong at SS/2B.
6/5/2012 10:42 AM
deanod
Posts: 3152 (0)
Hall of Famer
115 18 1 119 1 52
Prospect
Posted by dw172300 on 6/5/2012 1:01:00 AM (view original):
I think you guys are missing my point. I'm looking to put a value to the defensive stats beyond those metrics themselves. Has anyone used the defensive stats available in the game to calculate VORP or WAR? I've done it with offensive numbers but not the defensive ones.

I didn't miss your point- as kahrtmen noted defensive VORP can be approximated by (plus plays - (minus plays + errors))*2/3 - (whatever quantity you deem to be replacement level)

6/5/2012 10:56 AM
dw172300
Posts: 1173 (0)
Hall of Famer
1 182 5 92
Rookie
Posted by deanod on 6/5/2012 10:56:00 AM (view original):
Posted by dw172300 on 6/5/2012 1:01:00 AM (view original):
I think you guys are missing my point. I'm looking to put a value to the defensive stats beyond those metrics themselves. Has anyone used the defensive stats available in the game to calculate VORP or WAR? I've done it with offensive numbers but not the defensive ones.

I didn't miss your point- as kahrtmen noted defensive VORP can be approximated by (plus plays - (minus plays + errors))*2/3 - (whatever quantity you deem to be replacement level)

Apparently *I* missed *your* point.

Thanks. I appreciate you leading me a little further down the path on this. 
6/5/2012 2:52 PM
dahsdebater
Posts: 10046 (3)
Hall of Famer
284 293 18
Pro
For player comparisons the easiest thing to do might be to calculate an equivalent line.  IE for infielders, for whom it's a close approximation to say that all +/- plays and errors are roughly 1 base, add 1 single to the batting line for every plus play and subtract one for every - play and every error.  It's easier to look at an offensive line and see what it's worth/compare it to somebody else's.  Making equivalent lines for 2 guys lets you compare what they do on offense and defense without having to calculate WAR or VORP.

EDIT: for outfielders it's a little tougher, obviously.  You'll only add a fraction of a hit to the batting average for each + play but somewhat more than 1 total base to the slugging percentage.  Similarly, you have to work out what fractions to use for - plays and errors, but there are some good estimates around if you look for them.  Some SLB owners have done pretty good analyses to estimate the values of defensive plays (+/-/E) by position that probably map over exactly to HBD.

6/5/2012 3:59 PM (edited)
MikeT23
Posts: 58012 (0)
Hall of Famer
87 1 47 1 189
Prospect
That's the other thing I did to figure out a rock with a big bat is a good play in RF.    The best RF, a GG-caliber 3B, wasn't going to save more than 20someodd bases in RF over a DH-like slug.   Add 25 bases to the fielder's OPS and see what you get. 
6/5/2012 4:01 PM
litigator01
Posts: 18 (0)
Hall of Famer
12 40
Prospect

Don't I remember reading that you found defense matters a lot more with LF?  So the recommendation would be to essentially flip conventional wisdom, and use RF for a 40-40-40-40 big-bat guy, and keep a 700 OPS 70-70-70-70 guy in LF?

6/8/2012 3:51 PM
MikeT23
Posts: 58012 (0)
Hall of Famer
87 1 47 1 189
Prospect
Yeah, I said as much in The Rightfielder thread.    You can pick up a lot of + plays with those 85/65 players that kill you in CF by sticking them in LF.
6/8/2012 4:22 PM
jtrinsey
Posts: 237 (0)
All Star
1 4 2 30
Rookie
As a poster mentioned, every "extra" out your fielder makes is probably worth somewhere around 2/3 of a run. And every out they don't make (that an average fielder would have made) costs you like 2/3 of a run, on average. Outfield outs probably worth more in both directions if you want to get that involved.

So do the calculations of your choice to figure out (on a rate basis) how many outs per chance that your guy will make more (or less) than the average player at his position and figure out how many chances per inning that position will get and how many innings he'll play. Multiply by 2/3 to get a FRAA (fielding runs above average) and throw that into your WAR calculation.

An alternative shortcut that doesn't involve lengthy statistical modeling would be to figure that the best fielders at their position will save about 10 runs per year and the worst will cost about 10 runs. Maybe a little more in either direction for a very high-chance position like SS and a little less in either direction for a very low-chance position like RF. So the 95/95/95/95 SS will save you an extra 10-15 runs per year and the 75/75/75/75 guy will probably cost you 10-15 runs. And scale based off that. That sort of logic will fail once you get into really extreme stuff (like playing a SS at 1B or vice versa), but can probably ballpark it with a reasonable degree of accuracy. 
6/8/2012 11:16 PM
12 Next ▸
Advanced Analytics - Defense Topic
△Top△Bottom

Search Criteria

Terms of Use Customer Support Privacy Statement

© 1999-2026 WhatIfSports.com, Inc. All rights reserved. WhatIfSports is a trademark of WhatIfSports.com, Inc. SimLeague, SimMatchup and iSimNow are trademarks or registered trademarks of Electronic Arts, Inc. Used under license. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.