ProSLG

by raysrule07 on Aug 20, 2007 6:28 PM EDT Comment 5 comments

For those of you who read raysbb.com(formerly raysbaseball.com), you'll notice the stat IsoSLG(or IsoP)used often to measure a player's power, especially for prospects.

IsoSLG is supposed to isolate SLG % from BA in order to more accurately depict a player's power independent from their contact abilities. However, I think it hurts players who have lower batting averages, because it implies a linear relationship between SLG and BA, when the relationship is more proportional. Take the following (fictional) example statline:

500 AB, 100 H, 20 2B, 5 3B, 15 HR, 175 TB, .200 AVG, .350 SLG, .150 IsoSLG

Now, if this same player were to have 40 more hits to raise his average to .280, using IsoSLG, you would predict that his SLG would rise the same amount, to .430. However, this is only true if all 40 additional hits the player would have hit would have been singles, adding 40 total bases. Using the percentage of each extra base hit(2B-20%, 3B-5%, HR-15%), over 40 AB's, he would have hit 24 singles, 8 doubles, 2 triples, and 6 home runs, for an increase of 70 total bases, to 245, making his SLG % .490, and his IsoSLG .210.

As you can see, the relationship between batting average and slugging percentage is proportional, not linear. I believe a more accurate representation of a players power would be a proportional number found by dividing SLG by AVG. For instance, in the example above, the player's IsoSLG for the first line was .150, and his proportional slugging(or ProSLG) was 1.750. For the second stat line, his IsoSLG was .210, and his ProSLG was 1.750, the same as the first stat line.

There are of course, flaws, such as if a player has a low BABIP, using ProSLG to predict a SLG should the BABIP return to normal levels would be inaccurate, as the number of home runs would not have increased, etc.

The 5 leaders in the D-Rays organization are as follows:

Player        Age    Level ProP    IsoP
M. Fields    22    A 1.927    .227
D. Mohr        31    AAA 1.913    .209
Q. Stewart    24    A 1.871    .229
S. Pedroza    23    A-Adv 1.861    .247
R. Royster    21    A 1.785    .255

and others of interest
E. Longoria    21 AA/AAA 1.747    .224
J. Pridie    24    AA/AAA 1.622        .184
J. Ruggiano    25    AAA 1.633    .193
R. Brignac    21    AA 1.675    .172
D. Jennings    21    A 1.476    .150
J. Jaso        24    AA 1.502    .158
R. Hughes    24    A-Adv/AA 1.559    .181
F. Perez    24    AA 1.570    .171

0 recs | Comment 5 comments

Story-email Email Printer Print

Comments

Truly excellent thinking

Player----- BA-----SLG----ProSLG
Pena -----0.263-----0.546----- 2.076045627
Dukes -----0.19-----0.391-----2.057894737
Upton -----0.313-----0.526-----1.680511182
Navarro -----0.204-----0.316-----1.549019608
Crawford-----0.316-----0.483-----1.528481013
Harris-----0.293-----0.425-----1.450511945
Young-----0.293-----0.411-----1.402730375
Iwamura-----0.285-----0.39-----1.368421053

It makes BA less of a priority which is the purpose of the Iso stats.

I fully intend to adapt ProPower / ProSLG over IsoP / IsoSLG.

by R.J. Anderson on Aug 20, 2007 6:33 PM EDT reply actions 0 recs

I am not so sure

Raysrule 07- If you take the stat line you have above and add those same 40 hits but with the extra base ratio at (2b-10%, 3b-0%,HR-10%)You then have a line of:

500 AB,140H,24 2b,5 3b,20 HR,234 TB
.280 AVG, .468 SA, ISO .188

His ProSLG is 1.67 whereas your hypothetical player has a ProSLG of 1.75 even though my hypothethetical has more doubles and more home runs he is being penalized because he hit more singles. Are you sure this is what you want to do?

by TomT on Aug 20, 2007 7:17 PM EDT reply actions 0 recs

Re: I am not so sure

I am not sure I follow you Tom, I used the % for doubles(20%), triples(5%), and home runs(15%) to extrapolate the numbers to 140 hits from 100, so I am not sure where the 10%, 0%, and 10% came from. Alsoo, my hypothetical player has 28 doubles, 7 triples, and 21 home runs, yours has 24 doubles, 5 triples, and 20 home runs, so your comment about your hypothetical having more 2B and HR than mine is confusing to me also.

However, I do understand that players are penalized for hitting singles. ProSLG is found by dividing SLG by AVG, which leaves us a ratio which is, essentially, the number of bases per hit, not per AB like slugging percentage by itself. So, a single would drag the ProSLG down, towards 1.000, as opposed to doubles and home runs, which bring it up. Singles do not connote power, as they can be bunt singles, infield singles, weak ground balls into the hole, etc. Doubles and triples are more likely to be line drive hits to a deep part of the field, and the ball must fall between or go past outfielders, a single can fall in front of them. And home runs, of course, must be hit very hard. So yes, players are penalized for hitting singles, but since ProSLG is intended to help measure a player's power, I don't see a reason to reward a player for hitting singles.

by raysrule07 on Aug 20, 2007 8:18 PM EDT up reply actions 0 recs

Re: ProSLG

Sorry. I guess I should have made my comparison more clear. I was comparing your initial hypothetical the player who hit .200/.350 with 20 2b, 5 3b 15 HR to my player who hit .280/.468 with 24 2b, 5 3b, 20 HR(actually that is a little more than a 10% increase in HR).

Anyway, your player has a higher ProSlg despite my player having more doubles and Home Runs. My problem with ProSLg is say a speedster like CC hits a ball deep in the hole at short and beats it out for a single he is penalized for getting the single. If Navarro hits the same ball and gets thrown out he is not penalized. This just does not seem fair to me.

by TomT on Aug 20, 2007 8:35 PM EDT reply actions 0 recs

Re: ProSLG

Ah, I see now what you are saying. And this is the part of the usefulness of ProSLG. If you were to initially look at each player's line, you would assume that your hypothetical player(I will call him player A) has more power than my hypothetical player(player B). However, ProSLG shows that player B has more raw power, but because player A has better contact skills, his power is put to better use.

In my original post, I put up the numbers for the top 5 in our organization in terms of ProSLG, and Matthew Fields was #1, while Ryan Royster was #5. This indicates to me that Matthew Fields has more raw power, but due to an advantage in other factors, such as contact skills, Royster puts up better power numbers. ProSLG is not intended to show what a player's power results will be or should be, but what they could be. If Matthew Fields were to improve his contact skills to Ryan Royster's level, I believe he would put up better power numbers.

As for the CC/Navi example, as I said in the original post, ProSLG isn't a perfect stats, and there are flaws such as this one. However, on the flip side, CC will also use his speed to stretch a hit to an extra base where Navi wouldn't, so I think it will even out pretty well, or even work to the advantage of the speedier player.

by raysrule07 on Aug 20, 2007 9:29 PM EDT up reply actions 0 recs