*For the interests of this discussion WAR will refer to Wins Above Average. It’s such a commonly accepted term that I will use it, but know that it does not follow the traditional form that looks at Wins Above Replacement.
Around this time last year I attempted to take a look at how each team fared by position. R/A!!! to @Radiohix for reminding me of this piece and motivating me to see it updated. The basic idea here is to compare the totality of each position to the league average using wOBA and UZR to get an idea of which teams were strong or weak at each of the positions on the diamond. You don’t have to understand all the lingo used here to appreciate the final results. With that in mind, I’ll show the final chart first and then show you how we arrived at this (Click all images to enlarge):
A lot more to follow after the jump…
If you read the previous piece, then this should make a lot of sense to you, but if you didn’t we can see how each team fielded a position above/below the baseline (0). We’ll cover the Rays, exclusively, in a bit, so how did some other teams work out? We can see that the Rangers LF had the most WAR of any other position on any other team with 5.7 WAR (makes sense since MVP Josh Hamilton played the most), while the Indians CF was the absolute worst position for all teams in the league at -4.2 WAR (this also makes sense since an injured and ineffective Grady Sizemore has not been the player that the Indians thought they were getting when they inked him to what seemed like a decent enough deal). There’s a lot more in between that will interest many, many people depending on who you root for.
So how did we get to this point? Well we start with the excellent B-Ref database that yields the splits for how teams did with the sticks by position:
Above, you will find the wOBA ((.72*(nIBB)+(.75*HBP)+(.9*1B)+(.92*ROE)+(1.24*2B)+(1.56*3B)+(1.95*HR)+(.25*SB)+(-.5*CS))/(PA-iBB)) for each of the leagues and the Rays. I love that for all of the talk about how great the AL is, it appears that batters in the NL were better hitters at all but RF and DH (unsurprisingly). I’m sure there are reasons beyond being better hitters, but it’s interesting, nonetheless. For those, like me, that are better with charts:
The Rays were extremely above average at 3B and LF, about average at 2B, SS (compared to AL), & CF and pretty well below average at C, 1B, RF, and DH. We didn’t hit horrible, compared to other positions, at 1B and RF, but you can see just how good the average team was at these offense-first positions. On the same note, you can see just how bad the C and SS positions hit (especially in the AL). Using these numbers, we can compare the Rays bats by position against the league averages pretty easily. We can clean this up a bit though, since it’s hard to convey how important (in terms of runs) these gaps can be. Also, we want to incorporate plate appearances since the more a batter steps up to the plate, the more his impact is felt. We can do this through a statistic called wRAA ((wOBArays-wOBAal)/1.15*PA) which shows how many offensive runs a player is above or below average. Here’s the Rays compared to the AL and MLB:
Our baseline is still 0 and you can see about what we saw before, but now with specifics regarding what these gaps mean. You can see that, compared to all of MLB, our 1B production was worth -13 runs (data labels correspond to wRAAmlb) and our 3B was worth 24.4 runs. Now that we have the number of runs that each position was worth, we can incorporate defense using a weighted average UZR of all the players that played at a specific position using the numbers at Fangraphs. An example could be the excellent 3B for the Rays: ((11.1*1330)+(-.1*10)+(-.5*10)+(-1.1*40)+(-1.8*25))/(1330+10+10+40+25) = 10.1 runs in the positive. Another example would be our less than excellent SS where we have a good player getting some of the innings and a bad player getting most of the innings: ((3*340)+(.3*9)+(-10.4*1104))/(340+9+1104) = -7.2. We can add this directly to the wRAA numbers and get an idea of the WAR for each position (this leaves out the replacement level and positional adjustments which is fine since we’re comparing to the league average for each position and is why I drew attention to this in the title). Pretty simple concept and the data is readily available if you don’t mind a little grunt work and yields this:
I have all of these charts for each team so if you are interested shoot me an e-mail at sabresrule79080ATyahooDOTcom and I’ll send you the entire workbook. I did all the work so that other people can just present the data to their own fans. Again, we can see that where the Rays were strong and weak and what the impact was, overall. Adding in defense sure did a number to SS, while boosting 3B and LF even higher. Additionally, 2B and CF checked in as a bit above average while 1B proved to be our weakest link. I think the initial chart now makes a lot more sense as you can see the above chart for each team for each position. Now let’s take a look at each position with it’s own scale as I think this will yield more useful information (though being less convenient than one master chart).
We can add all these up for each team and get this:
Wow, how bad was that Mariners team at the plate and in the field? They were nearly twice as bad as the second worst team, the Orioles. What I take out of all this is that it’s hard to replace an extremely glaring hole (as is the case with the White Sox having three very strong positions but got killed at second and third base. No surprise that the Yankees didn’t have a negative position and that the Mariners only had one bright spot. There’s a ton to cover for other teams, so I urge those that write passionately about their teams to shoot me an e-mail and break it down better than I can for your teams.
For those that prefer a table:
Lastly, I went through and ranked all the teams by each positions which you can see here:
I’ve broken this down about as well as I can, but would love to see where else we can take this (besides the obvious extension of looking at NL teams) so please give a shout if you think you have a good idea, and get in touch with me @Sandykazmir or at the above e-mail if you have any questions or want the data.