Steamer vs. ZiPS: Where the projection systems differ

Just before Christmas, FanGraphs released their ZiPS projections for the Seattle Mariners. Because of the holidays, we never really had a chance to really dig into these new projections. Colin briefly took a look at the over/under for these projections. I wanted to take a look at where the two major projection systems, Steamer and ZiPS, might disagree. Generally, Steamer is more optimistic while ZiPS is more pessimistic but individual projections can vary significantly from system to system.

I’ve pulled four players, two position players and two pitchers, whose projections are vastly different. I’ll dig into each to try and see why one system might be bullish or bearish

Ketel Marte

	PA	AVG	OBP	SLG	OPS	wOBA	WAR
Steamer	558	.269	.312	.356	.667	.293	1.8
ZiPS	585	.267	.308	.367	.675	.306	2.3

On the surface, the two projection systems see Ketel Marte producing a fairly similar batting line. Steamer thinks he’ll strikeout less often than ZiPS does but is bearish on his ability to hit for power. He won’t hit many dingers—that much the two systems agree on—but he’ll need to leverage his gap power and speed to reach the projected wOBA ZiPS is calling for. Despite a merely average defensive projection from ZiPS, it’s projecting an additional 0.5 zWAR in almost the same amount of plate appearances. If you think Marte is an above average fielder, ZiPS’s projection looks even better and indicates a significant amount of upside to the Mariners’ young shortstop. On the other hand, Steamer sees Marte as a plus defender and calls for a bearish WAR total anyway. A 13 point difference in wOBA may not seem like much but it’s the difference between Addison Russell and Elvis Andrus.

Leonys Martin

	PA	AVG	OBP	SLG	OPS	wOBA	WAR
Steamer	529	.242	.294	.350	.644	.282	1.2
ZiPS	446	.247	.295	.367	.662	.301	2.5

This is the most divisive projection on the offensive side of the projections. Despite fewer plate appearances, ZiPS projects more than twice the amount of WAR for Leonys Martin next year. Much of that is due to the difference in defensive projections. ZiPS thinks Martin is an elite defender in center field while Steamer thinks he’s merely above average. But the offensive projection is vastly different as well. Both projection systems see him bouncing back from an injury plagued 2015 but ZiPS sees him reaching the heights of his debut season where he was worth 2.9 fWAR. A higher projected BABIP and more robust slugging numbers help him get there but his primary value to the ballclub will be his ability to kill flying things in the outfield.

Steve Cishek

	IP	K/9	BB/9	HR/9	ERA	FIP	WAR
Steamer	65	7.88	3.28	0.97	3.85	3.91	0.2
ZiPS	61.6	9.19	3.35	0.73	3.50	3.14	0.4

In Cishek, we have another bounce-back candidate who has the second highest difference in FIP between the two projection systems. Steamer is pretty pessimistic, seeing last year’s struggles as his new talent level. Add a league average home run rate and Steamer’s projected FIP is higher than any other year in his career. ZiPS is much more optimistic and sees a return to his dominant ways. His projected walk rate is still a bit elevated but the projected FIP is the best in the bullpen. It comes down to how much you think his struggles last year were driven by injury or mechanical issues. ZiPS is drawing upon his much longer history of excellent performance and that’s the bet that Jerry Dipoto is making in signing him to be the closer.

James Paxton

	IP	K/9	BB/9	HR/9	ERA	FIP	WAR
Steamer	72	7.10	3.40	0.88	4.11	4.09	0.8
ZiPS	102	7.50	3.35	0.79	4.06	3.76	1

Despite spending the majority of the last two years on the disabled list, James Paxton has established a fairly consistent level of performance across the 141 innings pitched he’s accumulated. Both Steamer and ZiPS see Paxton dropping his walk rate lower than his career rate of 3.55 BB/9. Even in the minors, his walk rate was never this good. Overall, ZiPS has a more optimistic appraisal of his skill set, a better strikeout rate and a lower home run rate lead to a projected FIP that falls somewhere between his 2014 and 2015 marks. If a suppressed home run rate is truly a skill, like he showed to an extreme degree in 2014, then that will go a long way towards helping him reach his ZiPS projected FIP.

Win Projection

Neither Steamer nor ZiPS has a win projection built from their projections publically available. FanGraphs’s projected standings are currently offline until all of the ZiPS projections for every team have been released.

So, using the raw data from each of these systems, I built my own win projection for the Mariners. Using the BaseRuns formula, I took the projected data and calculated expected runs scored and expected runs allowed for the Mariners in 2016. Then I threw those numbers into the PythagenPat formula to come up with a win projection.

Note: ZiPS doesn’t produce projections that are based on projected playing time while Steamer produces projections that are based on the FanGraph’s depth charts. To build the win projection based on these two projection systems I scaled the ZiPS projections based on the projected playing time that Steamer uses. It’s not ideal but it’s easily implemented.

System	W	L	W%	RS/G	RA/G
Steamer	86	76	0.531	4.39	4.11
ZiPS	84	78	0.517	4.41	4.24

Overall, Steamer comes out looking more optimistic than ZiPS. It expects the Mariners’ pitching staff to be much more effective pushing their win total higher than the ZiPS projection. Offensively, ZiPS has a slight edge over Steamer, giving players like Nelson Cruz, Leonys Martin, and Ketel Marte the benefit of the doubt. At this point, 84-86 wins feels good enough to be in the conversation for a Wild Card berth, though it may only be good enough for third best in the division. As we’ve seen the last few years, teams who are able to stay around .500 for the first half of the year have been able to make some significant additions at the trade deadline to push them over the hump. That may be the destiny of the 2016 Mariners.

The offseason is far from over and there’s still a lot that can change. At this point, it’s reassuring to know that the two major projection systems see an above average roster, both offensively and in the pitching staff. They may disagree on some of the pieces, but overall they’re looking fairly optimistic.