FanPost

Projecting the Mariners

At the time of writing, the Seattle Mariners are 54-50 and 0.5 games back in the race for the second wild card spot. With rumors flying as to the upcoming trade deadline, the benefit of selling prospects to improve their odds varies heavily based on how view their current roster. Dave Cameron rang the warning bell for all the of the second wild card contenders, pointing out that these teams are vying for a one game playoff in Anaheim. Not such a great prize.

But. The Mariners haven't made the playoffs in 13 years and the prospect of a wild card spot has made this most exciting campaign in a decade. So how good are the Mariners? At 54-50, the Mariners have won at a clip of .519, but Fangraphs' projection models paint a very different picture of the team. Fangraphs' playoff odds page believes the Mariners are a .503 team going forward. Not so hot.

But. The Mariners' season-to-date-stats projection from Fangraphs is a lot more favorable, suggesting the true talent level of M's, based on CoolStandings's version of Pythagenpat adjusted for remaining schedule, is .557. To put that in context, it's the 4th best in baseball, behind the third-place Angels (.572) and 5th place Dodgers (.549). The Mariners, by this measure, are the both the 3rd best team in the AL West and the American League.

The CoolStandings model, like that of Baseball Prospectus, relies on base runs and a modified Bill James Pythagorean W-L. As best I understand it, this means that the Mariners, by base runs, have been ‘unlucky' to the tune of the gap between their projected rest of season W% and their current W%, for a loss of 0.038 W%. 3.8% of the M's 104 games already played comes out to 4 wins, so the Mariners' ‘true talent level' record would be 58-46. That's still a good margin worse than the Angels, but it's also well above all the other wild card contenders. By this model, the M's would be expected to win 86.1 games this year, beating out the next-best Blue Jays by almost 2 wins.

Conversely, by the Fangraphs Zips-Steamer projection system, the M's should win 83 games and come a game or so short of the Blue Jays. By these numbers, the M's still finish the as the best team not to make the playoffs.

There is a huge gap between these projections, coming out to about 8 wins over a full season and 5 over the roughly two-thirds of the season already played. The rest of this article is an amateur attempt to account for that gap, and to assess how good the Mariners might actually be.

The CoolStandings model relies on past performance and therefore absorbs some mathematical ‘luck.' For example, the Mariners' runs allowed this season has to have been impacted by a league high strand-rate. For that reason, the figures are probably overly optimistic of the M's chances.

But that isn't to say the M's look like a .503 ball club either. To account for the difference between the M's .519 W% and their .503 Zips-Steamer talent level, I looked at the Zips projections for the rest of the season, but in ignoring Steamer, this is only a rough guess at how the projection could be too pessimistic .016% of 104 games is about 2 wins, so that's the extent of the disagreement between reality and luck against the projections. For the remainder of the season, that .016% is worth about one win.

For example, Michael Saunders has been worth 1.7 fWAR this year in 219 PAs, a pace of approximately 4.2 fWAR over 550 PAs, but Zips says he'll be worth 0.7 fWAR over 161 PAs the rest of this year, a clip of 2.39 fWAR/550. If we say the M's have played two-thirds of their season already, that says that Saunders should've been worth 1.6 fWAR this year, and has overperformed to the tune of .1 of the M's wins.

Looking a little deeper, we can see that Zips doesn't punish Saunders playing time projection very much despite his injury trouble, in that it sees him basically playing full time the rest of the way. If we take the Zips projection and put it over his 219 PAs of 2014 service time, the picture is a little clearer: Zips says Saunders should've been worth 0.95 fWAR this year. So Saunders has already given the M's almost a full win more than he should've, says Zips.

Other contributors to the M's supposed over-performance include Felix and Ackley.

Ackley's defense has been fine so far, per UZR, but Zips says it should be bad. So despite underperforming at the plate by 9% per wRC+, Zips says Ackley's glove has already given the M's a full half-win more than can be expected.

A topic of some discussion, Ackley's fielding is hard to assess. To the naked eye he's looked ok, not terrible, and UZR seems to agree. There was his spectacular catch the other day, and there's also his mediocre arm, which possibly has been taxed at a below-average rate, but this is all speculation. But if we trust Ackley's to date figures and performance, we can give the Mariners a third of a win back over their remaining schedule.

And then there's Felix, who has already contributed a remarkable 5.5 fWAR. Zips penalizes Felix quite a bit, expecting him to regress quite a bit. If Felix continues pitching at his current level, the M's again would be expected to win another half game more than Zips suggests. Felix both has reached new highs with his changeup and has benefited quite a bit from Zunino's pitch framing, as have the M's, and Zunino's framing is neither accounted for by the projection systems nor his season-to-date fWAR.

Then there's Chris Young. Young has been worth 0.6 fWAR and 2.7 rWAR. Young alone can account for the gap between Fangraphs' and Zips' perception of the Mariners and their current performance. A lot has been written about Young's season, and going back to 2009, the last season Zips looks at for their projection's, Chris Young's 2014 is the single greatest overperformance of ERA against FIP. Young probably hasn't been as good as his ERA nor nearly as bad as FIP, but I can't speculate at his true talent.

Between Young, Zunino's framing, Ackley's defense, Saunders' somewhat expected improvement, and Felix's dominance, the Mariners seem like an especially tough team to project. For a counter example, the Blue Jays are expected to win 84.3 games by Zips-Steamer and 84.9 by Cool Standings, versus the Mariners 3 game swing between the projections. Clearly Zips-Steamer is the more reliable model and clearly it's missing a significant piece of the picture. But were I a betting man, I'd certainly bet the M's finish better than 83 wins.

What these numbers suggest is that by both the models the Mariners are close enough to be competitive for the wild card, and that acquiring marginal talents like Marlon Byrd or a DHing Matt Kemp (not going to happen) could have a real impact on the team's chances. By both models, going all in for Ben Zobrist at SS and a right-handed OF might not be be such a bad idea, nor would be so ludicrous to pursue David Price. That said, a bad trade is a bad trade no matter the context of standings, and the M's suffer from an overpay in any event.

What the holes in the Zips projections say, however, is that maybe the M's recent slide isn't especially important, and that while this probably isn't the 4th best squad in baseball, it's still probably a good team, and a team to be excited about. Because even if the M's fail to make another roster move, they should be a competitor.