Player OPS benefitted by his team's offense

[Triad]

The hypothesis here is that a team's offense contributes to a player's OPS. I'll present my findings, and then see what conclusions we can reach from them.

I took 24 players from before the juiced ball era who had full seasons for seven straight years from age 25-31, which could be thought of as the prime for most good players. From each player's seven-season sample, I ranked their own OPS's from best to worst, and then looked at any possible correlation for their teams' offensive outputs.

The players I used in the study:
Toby Harrah
Ralph Garr
Ron Cey
Richie Zisk
Rod Carew
Ken Singleton
Steve Garvey
Thuman Munson
John Mayberry
Al Oliver
Graig Nettles
Amos Otis
George Scott
Bob Watson
Tim Raines
Tim Wallach
George Bell
Gary Gaetti
Steve Sax
Julio Franco
Joe Carter
Kevin McReynolds
Andy Van Slyke
Kirby Puckett

Here's what I found...

For the players' highest OPS seasons, they occurred at the following distribution:
#1 Tm R/G ... 33%
#2 Tm R/G ..... 8%
#3 Tm R/G ... 25%
#4 Tm R/G ... 17%
#5 Tm R/G ..... 8%
#6 Tm R/G ..... 4%
#7 Tm R/G ..... 4%

"#1 Tm R/G" means it was the highest team run output in that player's 7-team sample. "#7 Tm R/G" means it was the lowest.

So, from this sample so far, it seems like there is a fairly strong positive correlation between how productive the team offense is and the individual player's OPS. It is heavily weighted toward the high-scoring years.

For each player's 2nd-highest OPS seasons, here was the breakdown:
#1 Tm R/G ..... 4%
#2 Tm R/G ... 13%
#3 Tm R/G ... 38%
#4 Tm R/G ... 17%
#5 Tm R/G ... 17%
#6 Tm R/G ..... 8%
#7 Tm R/G ..... 4%

Notice how it shifted down slightly, according to the team's output.

Now let's look at the next-to-lowest and lowest OPS seasons for each hitter.

Next-to-lowest OPS seasons for each player:
#1 Tm R/G ... 13%
#2 Tm R/G ..... 4%
#3 Tm R/G ... 17%
#4 Tm R/G ..... 4%
#5 Tm R/G ... 17%
#6 Tm R/G ... 17%
#7 Tm R/G ... 29%

Lowest OPS seasons for each player:
#1 Tm R/G ..... 4%
#2 Tm R/G ... 13%
#3 Tm R/G ..... 8%
#4 Tm R/G ... 13%
#5 Tm R/G ... 21%
#6 Tm R/G ... 25%
#7 Tm R/G ... 17%

So, when the team has a lower R/G output, the individual OPS's seem to suffer.

The variances in data for each player, comparing their top 3 averages with their bottom three averages, based on Tm R/G:

OPS R
.094 0.6
.053 1.0
.005 0.8
.036 0.5
.131 0.9
.102 0.6
.015 0.8
.011 0.8
-.012 0.9
-.004 0.2
.016 0.9
.085 0.7
.005 0.8
.064 0.5
-.015 0.6
.058 0.4
.077 0.5
.178 0.5
.057 0.4
-.049 0.6
.037 0.7
.065 0.7
.059 0.7
.025 0.4

Overall, this suggests that within the prime of the basic good hitter, a normal shift in R/G of his teams of 0.6 can be expected to produce an advantage of +.046 in OPS.

Incidentally, the distribution among different ages for the 1st and 2nd-best seasons of these players was as follows:

25 ... 21%
26 ..... 8%
27 ... 21%
28 ... 17%
29 ... 17%
30 ..... 8%
31 ..... 8%

What do you think we can derive from this study, if anything? Any questions about the methodology used here?

[kflo]

if everything around you stayed the same, your team would likely score the most runs in the season you had the highest ops. it would be interesting to look at either team ops or rc/g without the player being analyzed, and then looking at the correlation.

you also may have some selection bias. bill madlock and dave winfield would throw things off a little. i'm sure there are others.

also, some seasons (NL 1977 for example) have higher overall offensive production, which will inflate both r/g and ops.

not throwing out your analysis, because it's at least a start.

[nanwynnfan]

Just going by definition of terms, I'd suggest that OPS, which is the sum of OB% + SLG% would have little direct relationship to runs scored by a team.

The reason for this is that, in a game context and at the team level, the only direct effect game strategy, like the IBB, can have on a batter is an infrequent and minimal spurt in OB%. Conversely, the IBB reduces the player opportunity to up his SLG% by taking the bat out of his hands.

Since Runs Created can be calculated individually and at a team level, it might be more relevant to take a team's total bases and OB% with the individual batter deducted, to see the RC for the team without him. Then put him back in to determine his net impact on the team. That way you can relate the team to the individual player contribution.

Example: the Transylvania Silk Sox put up the following total numbers:

Total Bases............OB%..............Runs Created

2,620................ .341.................893

Dag Schmirnoff, a LF, puts up these numbers:

344................. .410.................141

Without Dag Schmirnoff the Sil Sox have:

2,276............... .330.................751

Actual figures would take team totals, subtract player, then check balance of totals with player put back in.

I believe RC, derived by TB*OB%, would make for a better situational interaction bewtween player and team.

[WilsonC]

It's a good start, but there are a few important adjustments needed to remove a selection bias from the results:
- As mentioned, the player's performance needs to be removed from the team's performance. If a player has a career year, he'll naturally create more runs, and weight the team scoring upward, when the hitter's performance is really the cause, not an effect of the team's run total.
- Consider also using an adjusted metric, like OPS+, or even something like EQA. The reason here is that a shift in the run-scoring context - moving the fences in, shifts in the league's pitcher/hitter balance, etc. will often affect both the hitter and the rest of the team - thus creating a cause for both the hitter's elevated offense and the team's offense.