clock menu more-arrow no yes mobile

Filed under:

Jason Marquis: Where Stats Disagree

Jason Marquis leads the team in wins and innings, but when looking at some of his 2013 advanced statistics there is a conflict brewing as to whether he's been an asset to the Padres or a detriment.


When it comes to basic statistics, it is pretty easy to just assume that Jason Marquis has been the Padres' best workhorse on the staff. He is 5-2 with a 3.49 ERA. Having watched Marquis, I think most would say those numbers overrate him maybe a little bit. We know he plays in Petco Park, a pitcher's park, and those that watch closely will have noticed that he has been putting himself in a lot of jams by leading the league in walks allowed. So, maybe some advanced stats are needed to really dig in to what the Padres are getting from Jason Marquis. However, that is where this get complicated.

Anybody who does not automatically tune out when advanced statistics come up is familiar with Wins Above Replacement (WAR). WAR is useful as a single number that will show the value of the player. Finding that value is the hard part and there are multiple methods. Two major baseball statistic websites use different methods. And with those different methods come two different evaluations of Jason Marquis' 2013 season.

Fangraphs: -0.7 WAR

Baseball-Reference: 0.4 WAR

So, one has put Jason Marquis under the microscope and concluded the Padres would be better off grabbing a pitcher from AAA and replacing him and the other says that he's doing an okay job. This is the Padres' main free agent acquisition of the offseason. I want to know what to make of him and this stuff is not helping. Why not? And what can we learn by figuring out why they disagree?

The reason why comes down to the methodologies. Fangraphs uses Fielder Independent Pitching (FIP), which is an ERA-like number that derives from what are called pitching component stats. Stats like HR/9, BB/9, K/9. Marquis does not do well in these areas. He gives up 1.5 HR per 9 innings, he is walking a career high 4.4 per 9 and is striking out a respectable, but not dominant, 5.3 per 9. FIP puts that all together and gives an ERA-like number of 5.62. A 5.62 ERA would certainly scream that he is hurting the team out there.

The Baseball-Reference WAR is based on Runs Allowed per 9 innings. Unearned runs count the same as earned runs since you do have to play with the defense that's behind you. The stat is adjusted for defense later.

The problem I see with Fangraphs methodology is that it is a fantasy and maybe Marquis is better than the fantasy. It assumes this robot of a pitcher that is consistent putting up the same performance based on the component numbers every time out. There are no good days and there are no bad days. It turns Marquis into this pitching machine.

If you break down Marquis' starts you get 2 absolutely brilliant games where he carried the team to victory with his pitching, one where he pretty much singlehandedly lost it for us (I was at this game, ugh), one start where almost every hit he gave up turned into a run (including 3 unearned runs) even though he was stingy with the walks, and then four games where he was walking a bit too many, striking out too few and letting the balls land where they may and got some luck.

That combination of outings does not say to me that Marquis has been awful. It does not say that he has been particularly good either, but I am not trying to build an All Star case here. Merely that when you break down Marquis' starts he does not come off as a pitcher that needs to be whisked out of the rotation. In fact, I kind of like an inconsistent pitcher that can be brilliant on occasions, but put up a couple stinkers on bad days. Better than the robot that is consistently being not-that-great. For example, if FIP is right and Marquis is pitching like a 5 ERA guy, then I'd rather if he pitches in 3 games give up 3, 8 and 4 runs for a 5 run average, than give up 5 in all three. If the team scores 4 a game then they might win one, lose one and compete in the 3rd rather than lose all 3.

It would seem that we find that his value is probably closer to what Baseball-Reference puts out, which is line with how I usually favor their WAR over Fangraphs'. However, even Baseball -Reference's methodology is somewhat flawed because if a pitcher allows a lot of balls in play and a lot of those end up as hard groundballs with cataracts or some hard liners that go straight to fielders or other such luck events, then you could hide that trouble is around the bend. Alerting fans to future struggles is what FIP is best at doing despite its flaws at trying to describe a small sample of past starts. So, Marquis needs to cut down on the walks or else those Fangraphs stats might turn from fantasy to reality.