It looks like it has. They seem to be scoring more runs than their OBP and SLG would normally indicate. The following equation shows the relationship between team runs per game and OBP & SLG for all teams from 2007-09:
R/G = 11.595*SLG + 16.04*OBP - 5.52
The Rays have a .338 OBP & a .410 SLG. That predicts 4.66 runs per game yet they are actually scoring 5.11. That is .45 more than expected and the next highest positive differential is about .30 for the Padres
I then used an equation which included GDPs, SBs and CSs per game. It was
R/G = -0.0556*GDP - 0.182*CS + 0.105*SB + 11.19*OBP + 16.34*SLG - 5.43
That predicted that the Rays would score 4.71 runs per game, still well below their rate of 5.11. Then I added in 4 other baserunning variables: the % of runners on first who make it to third on a single (13%), bases taken, like on fly balls and wild pitches (BT), reaching on errors (ROE), and outs on base, like getting thrown out trying for an extra base (data from Baseball Reference). The last three were all per game.
Here is the equation:
R/G = -.025*GDP - .347*CS + .083*SB + 11*SLG + 14.59*OBP + 1.17*13% + .359*BT - .426*OOB + . 699*ROE - 5.48
Plugging in all of the Rays data would predict 4.87 runs per game. That jumps us alot closer to the 5.11. Adding in all of the baserunning data closes almost half the original gap of .45 between their actual runs and predicted runs.
I also tried breaking down OBP & SLG into cases of none on and runners on base (ROB). Here is the equation:
R/G = 7.83*ROBSLG + 9.26*ROBOBP + 3.66NONESLG + 8.53*NONEOBP - 6.12
Notice how SLG is twice as important with runners on than with none on. With none on, the Rays have an OBP & SLG of .322 & .396. With ROB, they have .357 & .429. Plugging all that in to the equation predicts 4.74 runs per game. That is .08 higher than the very first equation reported (4.65). So if we combined that with the findings from baserunning, we would probably get something over 4.87 and we have a good idea why the Rays are scoring so many runs this year.
I use the runs created formula with and without SB/CS. With it, the Rays score 4.81 runs/game. Without SB and CS, they score 4.75 runs per game. So the boost due to the running game is small. The big boost comes from great hitting with runners in scoring position.
ReplyDeleteVery interesting. I decided to check out B-Pro's EQBRR just to see how it compared to what you have found. The Rays, as a team, are +13.2 runs. EQBRR doesn't include GIDP or ROE, but it should include the others you have here. If you add those 13 runs into the OBP and SLG estimated runs you end up at 4.75 runs per game, which is right about you got, but it ignores GIDP and ROE. If we use team wOBA, we get about 4.8 runs per game. Fangraphs includes SB and CS as well as ROE. It doesn't include anything else related to baserunning. Right? If we add those 13 baserunning runs from EQBRR we then get 4.93 runs per game.
ReplyDeleteMight the additional runs have something to do with park factors? What else could it be?
By the way, don't quote my numbers. It's late and I did these as quick as possible. I can double check them tomorrow. I might use some of your formulas here to look at others teams and compare it to what we get if we use wOBA and certain parts of eqbrr. I just realized that since wOBA on fangraphs includes SB and CS, I need to take that out. Tampa Bay has 3.7 Stolen Base runs according to B-Pro.
Maybe the additional runs are something as simple as the Rays hitting an unusual number of home runs or extra base hits with men on base?
I guess David answered my question. Thanks.
ReplyDeleteDavid
ReplyDeleteIn a way I agree with you and in another way I don't. When I added in SB, CS and GDP, I got about the same increase in runs per game as you did. But when I added in the other baserunning variables, the predicted runs took a bigger jump. Also, although I did not look at RSIP hitting, when I looked at ROB hitting, and broke down OBP & SLG into none on and ROB cases, the predicted runs did not go up that much. The Rays only have a .265 AVG with RISP this year. Their OBP & SLG are .367 and .424, about what they are with ROB. It looks to me that baserunning might play a bigger role than situational hitting.
Cy