Some interesting homebrewed WAR results

I averaged a bunch of different Marlins players’ component statistics and used Baseball-Reference.com’s Park Factors for Landshark Stadium and decided I’d share some of the results. I know, it seems like a lame idea for a post, but some of these results may not mesh with our observation of the players involved, so I figured some of them would be interesting.

Let’s start off with some pitchers. For this I took the FIP and tRA values found on FanGraphs, adjusted FIP to the runs scored scale, averaged FIP/.92 (the estimate scaling factor) and tRA, and used Pythagenpat to figure out the wins. Remember, according to the definition laid out by Tom Tango here, a replacement level starter is a 38% win% pitcher via Pythagenpat, while a replacement level reliever is a 47% win% pitcher. We could also adjust for league, but I’m doing that by using the league average numbers for runs allowed. Keep in mind also that I extracted the runs scored from interleague games (or at least I think I did, using B-R’s splits data for the league).

Josh Johnson: 5.2 WAR
Ricky Nolasco: 3.0 WAR

It’s no surprise those two have racked up the most wins for the team. I know it’s hard for people to think of a pitcher with an ERA north of 5 as a 3 WAR starter, but given Ricky’s peripherals, it’s more than likely he’s just run into a long string of bad luck.

Chris Volstad: 0 WAR
Andrew Miller: 0.9 WAR
Sean West: 0.9 WAR
Anibal Sanchez: 0.5 WAR

These are the remaining four pitchers that have logged significant innings as starters this season. You can tell why the team’s rotation has been a weakness so far this year. Volstad has given up too many homers, the rest of the pitchers have walked too many guys for their strikeout totals. Of the four, Miller has surprisingly recorded the best numbers, mostly because he’s not given up the long ball. His overall tRA is much harsher on him than FIP/.92.

Kiko Calero: 1.6 WAR
Leo Nunez: -0.6 WAR

Not sure why Nunez is still closing for the team.

Renyel Pinto: 0.2 WAR

I tend to disagree, but tRA likes him more because of all that weak contact he induces. Pinto has a consistently low LD%, down to 14.3% this season.

Let’s take a look at some hitters. For hitters, I used StatCorner’s wOBA and FanGraphs’ WPA/LI (game-state specific linear weights, in other words) as the offensive inputs. For defense, I used FanGraphs’ UZR. For baserunning, I’ll use BP’s Equivalent Baserunning Runs. For adjustment runs, I’ll use whatever’s listed on FanGraphs. Here’s our top 5 position players:

1. Hanley Ramirez: 8.3 WAR
2. Dan Uggla: 3.1 WAR
3. John Baker: 2.4 WAR
4. Jorge Cantu: 2.0 WAR
5. Cody Ross: 1.7 WAR

Hanley got even better, Uggla and Baker still look strong, and Cody took a hit thanks to a low WPA/LI compared to his wOBA runs above average. Cantu was an interesting case, as there was a discrepancy about nine runs on offense between WPA/LI and wOBA runs (WPA/LI liked him better). As a result, he came out to almost 2 WAR while FanGraphs has him at 1 WAR. Guess he really reached on error a lot. And of course, bringing up the rear:

Jeremy Hermida: 0.1 WAR
Emilio Bonifacio: 0.2 WAR

These two have been terrible. Did you know that according to BP’s EqBRR, Hermida is dead last on the team, coming in at -4 runs on the basepaths? There’s one thing where our eyes agree with the numbers: Jeremy Hermida is a terrible baserunner.

(Another fun fact: Nick Johnson is second to last on the team with -2 runs. He’s barely been here. Funny stuff)

Keep some of these numbers in mind (and of course, know that these are approximations, though at least they were done with good procedure) next time you debate the worth of a player. This should not be the only thing you consider, but they could be useful.