Rangers Team Win Values
Yesterday, we talked about team win value totals, and how they won’t always match up with pythag win projections due to their context neutral nature. Today, I want to look at an example of why this is useful information, and how conclusions about a team’s true talent level based on RS/RA can be faulty.
Let’s talk about the 2008 Texas Rangers. They won 79 games, finishing nearly at a .500 record, despite the fact that they were outscored 967 to 901. Using the pythagenpat formula, you’d get 75 projected wins, so an analysis based on pythag might say the Rangers were lucky to win 79 games based on how they played. However, the total win values of their roster for 2008 paints them as an 84 win team. That’s significantly better than both their pythag and their actual record.
The difference, as usual, is situational performance. The Rangers allowed 98 more runs than the second to last AL team in run prevention, the Baltimore Orioles. Even when you include a park factor to make up for the environment they play in, they were still clearly the easiest team in the league to score runs against.
Now, it’s true, their pitching wasn’t very good. But it wasn’t far and away the worst in the league either. Their 4.83 FIP is actually substantially better than the Orioles 5.14 mark. Yes, Baltimore had a better defense and played in a less offensive-friendly park, but those don’t explain a 100 run difference in runs allowed when Texas’ pitchers posted a FIP of .3 runs better.
The difference can be found in situational pitching. The Rangers were unbelievably terrible at stranding runners – their LOB% of 65.7% was 3% worse than the next worst team, the Colorado Rockies. Look at these situational lines.
Bases Empty: .766 OPS, 5% below average
Men On Base: .874 OPS, 12% below average
Runners in Scoring Position: .891 OPS, 13% below average
Bases Loaded: .878 OPS, 8% below average
With no one on base, the Rangers were just not very good. Put a runner on, though, and they become disastrously terrible. Rallies just piled on top of rallies, and the runs came through like a flood.
There’s good news for Texas fans in this, though. FIP is more predictive from year to year than situational performance, so while their inability to leave runners on base hurt them in 2008, we shouldn’t expect that to repeat itself in 2009. They’re still not likely to be good at stranding runners (their pitchers aren’t really good at anything else, either), but they likely won’t be historically bad again.
This is one of the things that the team win values can highlight for us. The Rangers might have given up a lot of runs, but the way they went about giving up those runs shouldn’t necessarily be viewed as an indictment on the future abilities of their pitching staff. With a better expected situational performance, we can expect Texas to do a better job of keeping runs off the board even if their pitching and defense don’t improve at all.
Print This Post
Am I correct in assuming that this phenomenon is not predictive? There would be no way for us to have known this about the Rangers in April or May would there?
We could have projected the Rangers to be below average at runner stranding, but yea, there was no way to see a 65.7% strand rate coming. Even if we knew their pitching and defense were both poor, the worst we could have projected was something like 68%.
What is the league avg strand rate?
And how is it possible to predict with any sort of accuracy a teams strand rate if concepts like “clutch hitting” are mostly attributed to luck. If you knew the Rangers had a 65.7% strand rate going into the last month of the season would you assume that it holds or regresses to the mean in the final month? Thanks for the help in understanding this work…
About 71%.
You’d regress back towards a mean. What that mean should be is still up for discussion.
Are you using 57 team wins as a replacement level to arrive at the Rangers 84 win projection?
Also why is there such a disparity between position adjustments from team to team? I see a clear DH effect, but why do certain NL teams have a 10 run spread?
It’s 47 wins, not 57.
So how do you get to 84 wins from 26.7 value wins?
You’re only looking at Position Player wins. Pitching adds another 11.9.
You’re not counting their pitcher wins. You have to look at both their position player wins (the 26.7 wins) and the pitcher wins (the 11.9 wins) to get 38.6 wins.
Bigcpa,
26.7 is only the Rangers offensive/defensive value wins. you also have to add in 11.9 for the pitching.
I would guess that the Rangers’ poor 2008 defense would slightly lower the expected LOB% but that wouldn’t account for this much of an outlier. Is that right?
Right.
LOB% is largely indicative of the effectiveness of a team’s bullpen, correct? What’s strange to me is that the Rangers and Rays had similar WAR bullpen values (~3 wins) in 2008, yet the Rangers’ LOB% was ~65% and the Rays were ~74%. Although the Rays’ defense was significantly better than the Rangers, would it really account for that much of a difference on its own?
What’s the explanation for that?
No, LOB% is not mostly about a team’s bullpen. Starting pitchers can strand runners too, and they pitch significantly more innings than relievers. This isn’t measuring how many batters a reliever strands when he comes in. This is measuring what percentage of baserunners a team leaves on base.
Dave — I’m curious as to how an at least somewhat predictable team tendency to strand runners ties into the oft-made statement that from an individual batter perspective there is no such thing as ‘clutchness’. What am I missing?
Why is it that the position adjustments for teams don’t add up to 0? If you had a team of league average fielders everywhere it would look like this (assuming each year is 150 games):
pos UZR adjustment
c 0 12.5
1b 0 -12.5
2b 0 2.5
ss 0 7.5
3b 0 2.5
lf 0 -7.5
cf 0 2.5
rf 0 -7.5
dh 0 -12.5
total defense is -12.5 (for an AL team). So actually the AL teams wouldn’t be zero, but this doesn’t explain the variance between teams. Shouldn’t all the 1b for any team have a total positional adjustment of -12.5 in 150 games? Am I missing something?