Top 10 Worst contracts of 2000s - Suppan not on list

[CheezWizHed]

http://sports.yahoo.com/m...0TjERvLYF?urn=mlb,206780

 

All I can say is WOW for the LA Dodgers... 3 of the top 5! If that had been the Brewers, they would have been contracted.

[Invader3K]

Surprised Suppan isn't on the list. I guess maybe it's the fact that if it was a larger market team, the size of his contract wouldn't have been that big of a deal. Plus he was at least decent his first season.

[joepepsi]

At least Zito, Gary Matthews, and Vernon Wells have been useful, if overpaid, players.

[Ennder]

All of those deals were worse than Suppan's as was Pierre's which wasn't even listed. He actually earned his money the first season, the second season he was overpaid but not excessively so(no I don't think pitching WAR is useful at all, it is horribly flawed~). Last year was pretty bad though and I imagine this year will be downright atrocious since he got a raise and his production is going the wrong way.

[RobertR]

Alfonso Soriano is going to join that list right soon. At least the 2005 to 2015 list.

 

BTW, I disagree that Gary Matthews has been useful. I'd take Suppan over him even with an ugly 2009.

 

Robert

[joepepsi]

BTW, I disagree that Gary Matthews has been useful. I'd take Suppan over him even with an ugly 2009.

 

Robert

You're right. Matthews has been horrible with the Angels.

[logan82]

Last year was pretty bad though and I imagine this year will be downright atrocious since he got a raise and his production is going the wrong way.

 

Suppan is getting paid the same in 2010 as 2009 so I am not sure what raise you are talking about unless you are referring to the $2M buyout.

[rluzinski]

How is pitching WAR horribly flawed? Is it the 5.5ish ERA replacement level? The Brewer's 2019 season could be a great case study of how a 5.0 ERA pitcher still has positive value.

[Ennder]

Nope I was just mistaken logan, thought his salary went up this season.

 

Assuming you are pulling WAR from fangraphs they use FIP to build it which is a pretty flawed stat in the first place since it overcorrects for things like BABIP and LOB% and completely ignores HR/FB%. They need to rebuild how they figure WAR for it to really be useful. In fact I'd argue the stat doesn't even know what it is trying to be. It doesn't do a great job of showing how good or bad a pitcher is, it certainly doesn't give a good idea of how good his results are since they correct it for BABIP/LOB% but not for other factors. It is just some hybrid stat that doesn't really tell you anything except how good a FIP a pitcher had.

 

For hitters it is pretty good, though I don't trust the defensive side of it all that much and I tend to think park adjustments are handled poorly and muck things up some. I mean Dodger stadium is a better park for Manny to hit in than Fenway was but I'm sure he gets bonus WAR for being in a 'pitchers park'. I'm sure Juan Pierre does too and it really shouldn't be hurting him at all.

[endaround]

He doesn't like it because it includes HRs as part of the pitchers talent instead of ignoring them like xFIP does.

[Ennder]

He doesn't like it because it includes HRs as part of the pitchers talent instead of ignoring them like xFIP does.

I don't love xFIP either but it is certainly better than FIP or just looking at ERA. A pitcher has some control over his BABIP, LOB% and HR/FB% but all 3 stats are very high variance so you can't not correct for them at all, they should be regressing them towards his career rates, not just completely ignoring them and assuming all pitchers are the same for each stat.

[rluzinski]

If Fangraph's WAR calculation (didn't initially realize you were specifically criticizing their WAR calculation) were trying to estimate true talent, they should not simply use yearly HR rates. That's not what they are doing. They are estimating the context neutral value for a given performance. So Mauer might hit more HRs in a given year than his true talent suggests and Looper might give up HRs.

 

The reason why FIP is used is because it's both context neutral and defense independent. The defensive part isn't really a problem for batters because they are going to roughly face average fielders over the couse of a year. I agree it's problematic to assume a league average BABIP for every pitcher. A UZR-based estimate for non-HR BIPs would be better, IMO. tRA doesn't know hit location (UZR does) but it does use the GB/LD/IFB/OFB designations for BIP. Depending on how much you trust those subjective designations, tRA might be a more attractive choice compared to FIP.

 

If assuming a league average LOB% bugs you, so should using the league average value of a HR for a batter (ignoring clutch hitting). I think it can be reasonably argued that pitchers and hitters have some context based control over performance but I'm not sure if it's worthwhile to consider it (the estimates control is generally small).

 

I do generally agree that the stat isn't sure what it is but I think it's true for both batters and pitchers. If it wants to measure a player's actual value to their team for a given year, why is it context neutral? Why not just use WPA, then?

 

If I'm doing a lazy projection, I'll often just average FIP and ERA. Sometimes I include tRA in that average. If I had to use just one, it would probaby be FIP.

[Ennder]

If assuming a league average LOB% bugs you, so should using the league average value of a HR for a batter (ignoring clutch hitting)

 

I think you either need to use league average for everything or for nothing. The fact that it corrects two aspects that are high variance and ignores the 3rd makes it flawed in my opinion. I would want to either use just ERA by itself and ignore the flaws with it or use xFIP which attempts to correct for everything. LOB% is no more random than HR/FB% is. If a pitcher gets unlucky with BABIP or LOB% his WAR is going to be higher than you'd think basically, if he gets unlucky with HR/FB% it doesn't change. That makes no sense to me at all, why correct 2 areas and not the 3rd.

 

In this specific case I do not think the fact that Suppan got lucky with HR/FB% in 2007 and a little unlucky in 2008 suggests he was way better in 2007 than in 2008. The FIP trend is 4.42/5.51, the xFIP is 4.90/4.79 and the WAR was 2.5/-0.2. The ERA trend was 4.62/4.96. To me that points to a flaw in the system more than anything, especially for an established pitcher like Suppan.

[rluzinski]

I'm still not understanding why you wouldn't use actual HR count for pitchers. Again, we are not trying to estimate talent, just performance. A pitcher gave up the HR by himself, without his defense (most likely) influencing it in any way. We just use the average value of that HR (just like we do for batters).

 

LOB% is no more random than HR/FB% is.

 

The way fangraphs defines its "value" metric, it doesn't care if something is random or not. It only cares about two things:

 

1. Is it context neutral? (we ignore base-out situation and instead use the average run value of batting events)

 

2. Over a season, could the aggregate performance have been influenced by a potentially biased external factor? (batters face an average defense, so no issue; in contrast, pitchers could possibly have a very good or bad defense).

 

Now, I think fangraph's approach yields a number that while not necessarily wrong, has little practical value. It just tells you what that performance was worth, had bought it on the open market. That's it. So it can tell you if a trade ended up working out for a team but not necessarily whether the GM made a mistake at the time of the signing. If a stud pitcher falls apart after being signed, the signing ended up being terrible for the team but might not have been bad at the time of its signing.