If You Remove…
As Dave noted early yesterday morning, the Joe Blanton-to-the-Phillies trade is one that fans of the Phightins—like me—are not going to be in favor of for years to come. It’s not that Blanton is a bad pitcher, or that a guy like Adrian Cardenas is a lock to become a superstar middle-infielder, but rather that the Phillies gave up two of their top prospects for a back-end of the rotation pitcher. From living in Philadelphia and watching the local broadcasts and such it seems to me that those in favor of this trade feel the way they do because Blanton is “new.” If these same fans watched Blanton intently over the last year and a half or so I have a funny feeling their dispositions would not be as sunny.
Add in the facts that Kentucky Joe dominated the Phillies lineup a couple of weeks ago, that he “won” 30 games in 2006-07, and that Adam Eaton stinks and it becomes very easy to see how fans could be manipulated into thinking this was the right trade to make.
One form of this manipulation really irked my family yesterday, however, in that some writers and reporters tried to justify the trade with the clause “if you remove Blanton’s X worst starts, his ERA is …” I’ve seen differing quotes, involving Blanton’s 3.87 ERA sans-three bad starts, or even his 3.58 ERA sans-four bad starts. Well, you just cannot do that. My brother, who covers the Phillies at MVN.com, offered this take, which is essentially the entire point of this post:
“Todd Zolecki wrote on Philly.com that Blanton actually has a 3.58 ERA if you take away his worst four starts. That’s great, but you don’t merely take away four starts to try to justify a trade. Heck, if you take away Adam Eaton’s worst four starts, his ERA would be 4.06! Has Eaton been a pitcher who deserves an ERA nearly in the 3’s this season? No. Has he been effective at all? No. So let’s not cherry-pick stats.”
You can’t simply remove numbers to justify anything. As has been discussed at length this season, the last three years or so should be quoted when discussing a player’s true talent level. If these same writers wish to justify the trade, perhaps it would serve them wise to note that Blanton’s current ERA is worse than his pre-season projection, and that he is expected to be in the 3.77 range over the remainder of the season… not that his numbers would be better if we take away the worst starts, or that he has performed well against NL East teams in a total of four or five starts in his career.
Overall, we don’t know if the Phillies are done dealing. Maybe they turn around and send the perhaps-overrated Carlos Carrasco and fellow top-tier prospect Antonio Bastardo for Erik Bedard, which would give them a rotation of: Cole Hamels, Erik Bedard, Jamie Moyer, Joe Blanton, and Kyle Kendrick, with Brett Myers also in the fold. From what I’ve come to learn, however, regarding Pat Gillick and his “approach” on his way out the door, it really does seem Blanton is supposed to be the prized acquisition to keep the Mets at bay.
But then again, maybe he IS what the Phillies need to keep the Mets at bay, because as many other Philadelphia columnists have noted, Blanton has pitched 15 scoreless innings against the Mets in his career… (sighs and hopes readers pick up on the sarcasm of small sample sizes and cherry-picked stats).
Eric is an accountant and statistical analyst from Philadelphia. He also covers the Phillies at Phillies Nation and can be found here on Twitter.
Well, there may be something to be said for picking off the worst few starts — does it really matter if in his worst games he gave up six runs or ten? Either way the team he’s pitching for will lose. More generally, looking at the distribution of the number of runs given up, and not just the average, might be a good idea. Perhaps you might prefer someone who gives up, in two consecutive starts, zero runs in eight innings and ten runs in four innings to the guy who gives up five runs in six innings twice. The first guy is almost certainly one win and one loss; the second guy is a bit harder to predict but both of those games are more likely to be a loss than a win. But both of them gave up ten runs in twelve innings, having the same effect on their ERA.
But if people are going to cherry-pick Blanton’s best starts, they’ve got to cherry-pick everybody else’s as well. For example, removing Hamels’ four worst starts (7 in 5.7, 6 in 4, and 5 in 7 (twice)) takes him from 50 runs in 142.7 IP to 27 runs in 119 IP, or in ERA from 3.15 to 2.04. Every pitcher has good days and bad days.
Now, it may be the case that some pitchers are more inconsistent than others. That seems like a mental thing — some of them might let a couple baserunners get to them where others don’t. But in the absence of evidence, this business of calculating ERA without “bad” starts is meaningless.
And feel free to interpret this entire comment as my own desperate attempt to convince myself that the Phillies can hold off the Mets (and the Marlins) this year.
Isabel, the key point you made is that we have to do the same thing for the entire landscape of pitchers. If we remove the worst three starts from all starters, and then re-calculate an ERA+ type of metric, Blanton is very likely to still be where he is today. We can’t just do it to one pitcher to make him look better while leaving everyone else the same.
And I’m a Phillies fan, too, so I’m all for convincing in any way, shape, or form… as long as it’s correct and not idiotic. The Adam Eaton point my brother made really struck me hard and hammered the point home. Sure, Blanton has beena 3.87 starter sans-three bad starts, but Eaton’s been at 4.06 sans-four bad starts.
Cool post, though I’m hoping that the Mariners won’t make the trade you proposed. Gillick did enough damage to the team on his way out.
Incidentally, what happened to the post you had before about the possible effects of participation in the home-run-derby on hitters’ post-ASG performance? I was interested in seeing the comparison between 2nd-half projections and their actual performance, but the post completely disappeared.
From a Mariners standpoint I wouldn’t make that trade either, but from the standpoint of a Phillies fan I Welcome it with OPEN ARMS!!!
Some of my numbers were incorrect in that post and I’m extending it to a longer piece about the fallacy and inherent selection bias. E-mail me to discuss further.
Gooz fraba
Chill rocket, gooz fraba.
The only rationale that I can see for the Blanton trade is that it helps the Phillies in the play-off race this year. That’s the difference between an established player and a prospect (or two). The latter figure to be more useful maybe next year (and thereafter).
Billy Beane and Oakland got the better of that trade, as usual. As a former Pittsburgher (fellow Pennsylvanian), I feel for you.
Tom, here’s the thing, though: Adrian Cardenas, rated the top 2B prospect, wasn’t going to go anywhere in Philly with Chase Utley here, so it wasn’t even as if his performance would help the Phillies at any point barring a position change, which isn’t too likely. So it’s not that they gave up these prospects, but rather that they got a somewhat disappointing return. As I mentioned, Blanton will likely produce for the Phillies. It’s not that he’s a bad pitcher, but rather that he isn’t what you would think Cardenas and Outman would bring in return.
I actually don’t mind the trade from the Phillies standpoint. Blanton isn’t a world beater, but you know what he’s going to give you . . . . innings and should keep the Phils in the game. He’s not a rental so moving Cardenas is fine, plus Adrian is still 2-3 years away and Utley isn’t leaving anytime soon anyway. Outman has been alright, but with Happ proving that he is a legit, Outman is no longer the closest LHP to being MLB ready.
Obviously it’s a steep price to pay, but you gotta give something of value up to get something in return.
Brendan, you definitely make good points, I just feel as a Phillies fan and baseball analyst that either a) the return for two top prospects was not worth Joe Blanton, b) there’s something about Outman that we don’t know about, or c) the fact that it’s known Cardenas is completely a trade chip and serves no other purpose to the Phillies hurt them because teams knew they would have to unload him.
The point of this post was really that if people are going to defend the trade or say it’s good for the Phillies, which is a completely fine standpoint if it can be backed up, do not cherrypick numbers like Zolecki and others did. As I mentioned before, if we want to remove Blanton’s 4 worst starts, we have to remove the four worst starts from all starting pitchers and then look at ERA… and at that point, even though Blanton would be down to 3.58, it would be same in terms of being proportionately higher than the rest.
Now, if they mentioned that, perhaps Blanton pitched well in 15 of his 19 starts and that his higher ERA is due to a lower percentage of starts, that’s something, because even though it isn’t a perfect analysis it does tell us that he has a higher percentage of quality games than his barometer suggests… but to drop his 4 worst starts and nobody elses just doesn’t fly.
Eric, I wasn’t a huge fan of the trade either, but once I got over my initial reaction of “we gave up Cardenas AND Outman for that?!” I realized it’s not so bad. As mentioned, Cardenas wasn’t going to be in a Major League Phillies uniform probably ever, and Outman wasn’t really deserving of the #4 rank in the Phillies’ system. He’s really overrated, I think. Regardless, the Phillies’ Minor League system is that good anyway, so being in the top-five isn’t really an accomplishment.
I think we gave up a bit too much but I’d rather have given up those two for Blanton than possibly Cardenas, Carrasco and/or Marson for A.J. Burnett or Erik Bedard. I’m vehemently against a Burnett or Bedard trade because of their injury histories and Bedard is really overrated.
That should read “the Phillies’ Minor League system isn’t that good anyway.” My apologies.
Bill, regardless of what we may think of the trade I think everyone can agree on, or should be conscious of the fact that you just cannot remove numbers from one person and not everyone else to make something look good.
It’s very easy to remove the worst few starts for one person but if we want to remove them from Blanton we have to do so from everyone else. And at that point his ERA would still be proportionately higher than others.
I wrote a little bit back about something Jamie Moyer said in an interview, that he likes to evaluate pitchers by removing the 3 best and 3 worst games and looking at all the rest. Though it wasn’t a supremely extended study, the initial look showed it didn’t make a big difference to the average game score or the proportion between the players.
Let’s hope he extends it to 22 straight scoreless against the Mets!
Eric, have you heard of the new stat at baseball prospectus called Flake? It ‘s basically a standard deviation stat that measures which pitchers are the most consistent from start to start and which ones have had the most variance. For example, because of CC Sabathia’s god awful april starts and then dominance since then, his flake rating is through the roof. Since it’s a new stat, it’s hard to determine what its significance is right now, if, say, pitchers are consistently consistent or vice-versa, but it helps to account for some fluky bad outings that may warp season stats (like Bronson Arroyo’s disastrous ten inning affair a few weeks ago)
Adam, I have heard of it but have not looked at it a whole lot. It sounds quite interesting though. This all goes back to the true talent level, though. I actually wrote a whole big article this weekend at StatSpeak discussing Sabathia’s awful April with regards to his true talent level and projection going forward and it honestly didn’t change a whole heck of a lot. He had a 3.25 FIP projected for this year, and after his four starts, which is a very small sample, his projection changed to 3.47 for the balance of the season. Since then he’s had a 2.32 FIP, giving him a 3.23 for the season to date (or before that CG against the Giants). Which means he was projected to go 3.25 this year, at the halfway point was 3.23, and was projected for a 3.21 over the remaining 14 or so starts.
A few bad outings may cause a pitcher to have a very high rate in certain areas, but it’s all about how it affects his true talent level, not how it affects a half-season of data which isn’t indicative of such a true talent level.
If you’re going to take a player’s 3 (or 4) worst starts out, it would only be fair to also take out this 3 best starts.
NickP–
Eric did that very thing at Statspeak about a week or two ago.
NickP, yeah, I did something like that, albeit not a huge study, based on something Jamie Moyer said in an interview. He said he likes to evaluate pitchers after removing their 3 best and 3 worst starts, to look at the larger bulk of performance. Using Game Scores it didn’t make too much of a difference, and since the difference wasn’t substantial it doesn’t seem like removing starts period is any better at evaluation or serving as a predictor.