I’ll go with this right up front: We need another player rating measurement,
and I need all your help to get one perfected.
I really dislike the Wins Above Replacement rating system. A few examples below will begin to tell why, although there are virtually unlimited cases where WAR is just plain head-scratching.
One thing that seems rather strange to me is lumping all players – pitchers and position players – into a single rating system, but we’ll worry about that later. The initial focus is on position players.
I understand WAR tries to take into account more than just hitting statistics, and I think we might be able to get to that, but my initial attempt is generally to build a hitting measurement tool. I have not spent much time learning the calculation(s), and what little I did look at it, it looks almost impossible to do as a general baseball fan with the stats available to me.
There are actually two Wins Above Replacement systems generally regarded as relevant today – the Baseball-Reference WAR, or bWAR, and Fangraphs WAR or fWAR. I typically quote the bWAR numbers for players, because I just got more used to the Baseball-Reference pages. The two numbers generally track with each number, but there are some differences. I will not argue with anyone who thinks fWAR is better than bWAR or vice versa. I like neither of them.
Some background examples before I get into building a better mousetrap.
Two players:
Player A:
|
Yrs |
Games |
PA |
H |
2B |
3B |
HR |
R |
RBI |
SB |
BB |
SO |
PlayerA |
17 |
2196 |
8664 |
1933 |
383 |
36 |
434 |
1204 |
1289 |
152 |
891 |
1748 |
PlayerB |
19 |
2372 |
9692 |
2452 |
498 |
39 |
353 |
1296 |
1391 |
195 |
661 |
1741 |
Player A had a lifetime batting average of .263; Player B had a lifetime batting average of .277
Player A had a lifetime OPS of .839; Player B had a lifetime OPS of .793.
A little bit of math: Player B had 519 more hits in his career; 115 more doubles, but 81 less homers; he also had 92 more runs scored and 102 more RBI (with less homers).
In the January 2022 Hall of Fame voting Player A received 41.1% of the vote. Player B received 5.3%.
Oh yeah, Player A had 10 career gold gloves, Player B had 9.
Player A has 62.7 career bWAR; Player B has 50.7.
Player A has 67.0 career fWAR; Player B has 43.0 - a massive difference.
I cannot figure out how Andrew Jones, player A gets so much more love and attention than Torii Hunter, player B.
Next, we’ll do a chart for five players who all had the
initials “BB”.
Player BB1 is Barry Bonds, the gold standard of modern day WAR. The other 4 will be revealed after the data is shown.
|
Yrs |
Games |
PA |
H |
2B |
3B |
HR |
R |
RBI |
SB |
BB |
SO |
BB1 |
22 |
2986 |
12606 |
2935 |
601 |
77 |
762 |
2227 |
1996 |
514 |
2558 |
1539 |
BB2 |
14 |
1849 |
8090 |
1886 |
302 |
66 |
332 |
1258 |
1024 |
461 |
914 |
1757 |
BB3 |
22 |
2517 |
10037 |
2715 |
498 |
49 |
174 |
1077 |
1208 |
183 |
450 |
453 |
BB4 |
17 |
2213 |
9545 |
2375 |
277 |
131 |
54 |
1359 |
578 |
558 |
1129 |
907 |
BB5 |
16 |
2113 |
8257 |
2010 |
408 |
61 |
287 |
1084 |
1173 |
45 |
912 |
1204 |
Now if we eliminate Barry Bonds as ‘other world’ who leads
in almost all regular statistical categories, the remaining four are somewhat
even in the length of career and other measurements. Obviously BB4 is a ‘top of the order’ speed
guy rather than a power hitter, and BB3 accumulated quite a few more hits and
doubles without either the home runs or the walks.
|
bWAR |
fWAR |
Player BB1: Barry
Bonds |
162.8 |
164.4 |
Player BB2: Bobby
Bonds: |
57.8 |
57.2 |
Player BB3: Bill
Buckner |
15.1 |
18.8 |
Player BB4: Brett
Butler |
49.7 |
42.2 |
Player BB5: Bobby
Bonilla |
30.2 |
32.7 |
Bobby Bonds and Buckner both began their careers in the late 60’s. Bonds was done in 1981 while Buckner played until 1990. Both Butler and Bonilla played primarily in the 80s and 90s.
I cannot figure out how Bobby Bonilla has 2x WAR of Buckner, while Butler has 3X and Bobby Bonds almost 4X. Buckner leads all of them in games, plate appearances, hits doubles, RBI, and K-rate (by far). Of the four, only one has a higher career batting average that Buckner, and that is Butler’s .290 to Buckner’s .289.
So, what do we do? We create a better system! And with everybody’s help we can replace WAR with PEACE (yeah, I need some help here, too . . (Player Evaluation And C____ E_______)
I started this off with something that seems to make sense to me, but could probably use some tweaks and polishing. Hear me out for now on this and then offer up any constructive criticism and improvements. This system does not try to define a ‘replacement level player’, although there will probably be a value that is arrived at that can be considered just that. For the purposes of these examples, all are well above replacement players.
The system will create a score which will range in the hundreds (thousands for this group) to a high of 15045.4 for Barry Bonds. Nobody can get a negative score - if you make it to the majors, you don't deserve to leave with less than zero!
While Bonds' score is not an absolute limit, it will be difficult to surpass it, although if Mike Trout can stay healthy and highly productive for another 12 or so years, he may have a shot.
My system:
CATEGORY |
Multiplier |
Number of Years with at least one plate appearance |
Times 100 |
Games Played |
Divided by 10 |
Plate Appearances |
Divided by 20 |
Hits |
Times 1 |
Doubles |
Times 2 |
Triples |
Times 3 |
Home Runs |
Times 4 |
Runs Scored |
Times 1 |
RBI |
Times 1 - # Home Runs |
Stolen Bases |
Times 1 |
Caught Stealing |
Times -1 |
Bases on Balls |
Times 1 |
Strikeouts |
Times 0.5 |
Grounded into Double Play |
Times -1 |
Hit by Pitch |
Times 1 |
Sacrifice Hits |
Times 1 |
Sacrifice Flies |
Times 1 |
League MVP | ? |
Gold Gloves |
? |
These are rather arbitrary multipliers which I cannot argue very hard for, but I do want to start out by giving credit for playing time. This will act as the base for the end result. At any rate, the calc is the same for everyone.
I originally added a multipier of 100 for each MVP and 50 for each Gold Glove won, but have taken that out now so this is a pure hitting score. (Put your thoughts on this in the comments if you have them.)
So, how do the players above compare in the new system?
|
|
Barry Bonds |
Barry Bonds |
Bobby Bonds |
Bobby Bonds |
Bill Buckner |
Bill Buckner |
Brett Butler |
Brett Butler |
Bobby Bonilla |
Bobby Bonilla |
|
|
Stats |
Score |
Stats |
Score |
Stats |
Score |
Stats |
Score |
Stats |
Score |
Yrs |
*100 |
22 |
2200 |
14 |
1400 |
22 |
2200 |
17 |
1700 |
16 |
1600 |
Games |
/10 |
2986 |
298.6 |
1849 |
184.9 |
2517 |
251.7 |
2213 |
221.3 |
2113 |
211.3 |
PA |
/20 |
12606 |
630.3 |
8090 |
404.5 |
10037 |
501.85 |
9545 |
477.25 |
8257 |
412.85 |
Hits |
*1 |
2935 |
1495 |
1886 |
1186 |
2715 |
1994 |
2375 |
1913 |
2010 |
1254 |
2B |
*2 |
601 |
1202 |
302 |
604 |
498 |
996 |
277 |
554 |
408 |
816 |
3B |
*3 |
77 |
231 |
66 |
198 |
49 |
147 |
131 |
393 |
61 |
183 |
HR |
*4 |
762 |
3048 |
332 |
1328 |
174 |
696 |
54 |
216 |
287 |
1148 |
Runs |
*1 |
2227 |
2227 |
1258 |
1258 |
1077 |
1077 |
1359 |
1359 |
1084 |
1084 |
RBI |
*1– HR |
1996 |
1234 |
1024 |
692 |
1208 |
1034 |
578 |
524 |
1173 |
886 |
SB |
*1 |
514 |
514 |
461 |
461 |
183 |
183 |
558 |
558 |
45 |
45 |
CS |
*-1 |
141 |
141 |
169 |
169 |
73 |
73 |
257 |
257 |
57 |
57 |
BB |
*1 |
2558 |
2558 |
914 |
914 |
450 |
450 |
1129 |
1129 |
912 |
912 |
K |
*-0.5 |
1539 |
-769.5 |
1757 |
-878.5 |
453 |
-226.5 |
907 |
-453.5 |
1204 |
-602 |
GIDP |
*-1 |
165 |
-165 |
107 |
-107 |
247 |
-247 |
62 |
-62 |
169 |
-169 |
HBP |
*1 |
106 |
106 |
53 |
53 |
42 |
42 |
38 |
38 |
28 |
28 |
S |
*1 |
4 |
4 |
16 |
16 |
47 |
47 |
147 |
147 |
5 |
5 |
SF |
*1 |
91 |
91 |
64 |
64 |
97 |
97 |
51 |
51 |
97 |
97 |
MVP |
? |
7 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
GG | ? |
8 |
0 |
3 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
TOTAL SCORE |
|
|
15045.4 |
|
7946.9 |
|
9316.1 |
|
9022.1 |
|
7968.15 |
bWAR |
|
|
162.7 |
|
57.8 |
|
15.1 |
|
49.7 |
|
30.2 |
fWAR |
|
|
164.4 |
|
57.2 |
|
18.8 |
|
42.2 |
|
32.7 |
As a hitter, Buckner’s career WAR falls well short. I also do not understand the quite inflated
WAR of Bobby Bonds.
Can we do better?
And lastly, for my other personal soapbox to help Mr. Hunter get some more respect, my system comparison to A. Jones:
|
|
Andrew Jones |
Andruw Jones |
Torii Hunter |
Torii Hunter |
|
|
Stats |
Score |
Stats |
Score |
Yrs |
*100 |
17 |
1700 |
19 |
1900 |
Games |
/10 |
2196 |
219.6 |
2372 |
237.2 |
PA |
/20 |
8664 |
433.2 |
9692 |
484.6 |
Hits |
*1 |
1933 |
1080 |
2452 |
1562 |
2B |
*2 |
383 |
766 |
498 |
996 |
3B |
*3 |
36 |
108 |
39 |
117 |
HR |
*4 |
434 |
1736 |
353 |
1412 |
Runs |
*1 |
1204 |
1204 |
1296 |
1296 |
RBI |
*1– HR |
1289 |
855 |
1391 |
1038 |
SB |
*1 |
152 |
152 |
195 |
195 |
CS |
*-1 |
59 |
59 |
99 |
99 |
BB |
*1 |
891 |
891 |
661 |
661 |
K |
*-0.5 |
1748 |
-874 |
1741 |
-870.5 |
GIDP |
*-1 |
192 |
-192 |
262 |
-262 |
HBP |
*1 |
97 |
97 |
97 |
97 |
S |
*1 |
6 |
6 |
6 |
6 |
SF |
*1 |
71 |
71 |
71 |
71 |
MVP |
? |
0 |
0 |
0 |
0 |
GG |
? |
10 |
0 |
9 |
0 |
TOTAL SCORE |
|
|
8311.8 |
|
9039.3 |
bWAR |
|
|
62.7 |
|
50.7 |
fWAR |
|
|
67.0 |
|
43.0 |
This is a very small subset of the possible examples - my
next look will be to compare Lou Whitaker (75.1 bWAR) to Jeff Kent (55.5 bWAR).
A couple final notes: First, I do understand that this does not take into account where they play on the field and the fact that outfielders and first basemen should have better power numbers than shortstops. Once the numbers are calculated, other groupings can be done. Perhaps there is a “Shortstop Standard” and an “Outfield Standard”. I have not developed that thinking yet.
Lastly, I am ignoring (for now) two other factors. One is what kinds of teams did the player play on (Andruw Jones generally played on better teams than Torii Hunter did), and what ballpark did the player play in? (The Colorado Rockies vs. Citi Field factors). Perhaps that can be added in at some point, but for now, neither is a factor in my scoring.
Please comment with your thoughts. Also, please keep personal thoughts about each player (and PEDs) out of this discussion.
There's some serious work here.
ReplyDeleteMe?
I'm statted out.
The problem is that many ballplayers contribute positive attributes in a variety of ways. Vince Coleman was a deadly base stealer, but had no power and wasn't much of a defensive player. Rey Ordonez could field with the best of them but couldn't do much else. Dave Kingman could hit a ball into orbit when he wasn't swinging and missing. He was fairly useless in the field. Some players hit for high average but no run production. Others struggle to get past the replacement level in batting average but are still critical to their team's success. My question is whether a single stat makes sense at all. It would seem to me if we rated every attribute individually, then added up the ratings you would have a composite picture that is fairly accurate. For simplification's sake, on a scale of 1-100, Coleman would get a 98 on base stealing, but then probably a 20 on power and a 55 on batting average. You can fill in the rest on fielding, throwing, etc. Then add all of these numbers together to get a new type of WAR.
ReplyDeleteYes,I understand what you are saying.. I'm thinking that while this is primarily an approach from the hitting perspective, adding in some points for # of years played folds in some of the other talents, simply because lesser hitters can add value and stick around longer than those that are true replacement players. It seems that this somewhat 'normalizes' differences between a Brett Butler type of player and a Bill Buckner type of player.
ReplyDeleteI actually changed my measurement yesterday to use a multiplier of 50 rather than 100 for number of years which seems to make more sense
ReplyDeleteUsing this measurement, Barry Bonds still leads the pack, but Rickey Henderson gets very close.
ReplyDeleteI will have to remember to read this when I have access to more than my little I phone. and then comment. Ambitious effort.
ReplyDelete