A feature that gets a lot of attention here, probably, is our playoff odds page. That page uses updated player projections and manually updated team depth charts to determine playoff probability and expected record. I look at the page probably two or three times a day, and though the numbers mean only so much, there’s no better way to get an idea of where a team truly stands. Current standings tell you about the now; projected standings tell you about the significance of the now.
A feature that gets a lot less attention here, probably, is our playoff odds page based on season-to-date performance. It uses the same depth-chart information, but instead of using player projections, it uses what players have already done. For example, in the former case, the Rockies are projected with a half-decent Charlie Blackmon. In the latter case, the Rockies are projected with a terrific Charlie Blackmon. It’s evident why the former page is preferred, but the latter page can serve a purpose, especially if you’re wondering about potential under- and over-achievers.
I thought it could be interesting to compare the two pages. We’ll leave the playoff odds alone — those get complicated, and they’re not what this is about. Both pages have projected rest-of-season winning percentages. With which teams do we see the greatest differences? Is this as predictable as it seems like it would be?
I made a chart, that I only realized after the fact is pretty unhelpful. Here it is anyway, because I invested literally several minutes:
Note the different axes. On the x-axis, projected win% based on season-to-date numbers. On the y-axis, projected win% based on projected numbers from ZiPS and Steamer. You can see agreement that the Astros suck. You can see agreement that the Tigers and A’s are good. But then, just how good? Let’s examine the same data in table form:
|Team||Win%, Projections||Win%, Season-to-Date||Difference|
In one column, projected win% based on projected numbers. In the next column, projected win% based on season-to-date numbers. In the last column, the latter subtracted from the former. Some large differences show up. Also, you get non-differences, like with the Royals, but the interesting teams aren’t the teams in the middle.
If you evaluated the Rangers only by what they’ve done so far, you’d see a pretty lousy team, particularly after the injury problems. But then, it’s a team with Prince Fielder, whose slugging percentage matches his OBP. It’s a team with Adrian Beltre, who to date has been worth 0.1 WAR. It’s a team where Elvis Andrus has been less valuable than Robinson Chirinos. You can continue on down the line, and you find players who’ve been underachieving.
Right after the Rangers, you see the Diamondbacks. The Diamondbacks have been known underachievers, and a big chunk of the difference in this table has to do with the Diamondbacks currently having a staff RA9-WAR of -2.0. The issues have mostly been the pitchers and Martin Prado, and the projections still think the team is half-decent. Maybe a third-decent, but, not disastrous.
Then you’ve got a streak of expected contenders, topped by the Dodgers. People have been waiting for the Dodgers to kick it into gear, and through a quarter of the season the outfield hasn’t been great outside of Yasiel Puig. The catchers have yet to do much of anything, with and without A.J. Ellis. You look at that team, though, and you figure it has to win, with that rotation. Note that, in a twist, while the projections like the Dodgers more, the projections also like Dee Gordon less.
At the other end of the table, some 2014 surprises. The Marlins are referred to in the headline. Based on season-to-date, they’re projected for baseball’s seventh-best rest-of-season record. Based on ZiPS and Steamer, they’re projected for baseball’s sixth-worst rest-of-season record. The projection systems are reluctant to buy the offense, while the season-to-date numbers like even a Jose Fernandez-less Marlins roster.
The Rockies have a slightly bigger difference than the Marlins do. This has little to do with pitching, and almost everything to do with the crop of position players. Troy Tulowitzki has been the best player on the planet. Blackmon’s already exceeded all expectations, and guys like Nolan Arenado, Justin Morneau, and Corey Dickerson have been overshadowed and overachieving. The Rockies have been a position-player juggernaut; projection systems foresee a slow-down.
And then there are the A’s. The A’s, who don’t have Jarrod Parker or A.J. Griffin. The A’s, who’ve found gold in Jesse Chavez. The A’s, who’ve featured an unbelievable catching tandem in Derek Norris and John Jaso. The projections do like the A’s — right now, they have baseball’s second-highest playoff odds. But the other version of the playoff-odds page loves the A’s, who have baseball’s best run differential by 40. The worst pitcher’s been Dan Straily, and he’s been sent to the minors. The general message here: the A’s have been playing probably too well. But they might just be freaks.
It’s a tricky thing to discuss, the difference between projected record and projected record based on season-to-date data. Sometimes, certainly, the projections can lag, because projections need some convincing to change their minds. If you’re a believer in the Marlins’ offense, you might believe more in the alternative playoff-odds page. If you’re a believer in the Rockies’ offense, you might believe more in the alternative playoff-odds page. The regular playoff-odds page is going to be more conservative, but in most cases that tends to be the proper approach. The season-to-date page will more closely reflect how fans do feel about their teams. The projections page will more closely reflect how fans should feel about their teams. We’re all prone to recency bias, but then we do sometimes spot a change before the projections do, so we’re not total idiots.
According to a playoff-odds page, the Marlins are one of the best teams in baseball. According to a playoff-odds page, the Marlins are one of the worst teams in baseball. Those are two facts. It’s up to you how you choose to weight them.
Print This Post