That's clearly not what I said, and glib to the point of being useless. And let's not be vague with terms like "good" or whatnot, which just leads down the road of pointless semantics. There are three issues at hand: Who was Ryan Dempster coming into this season, how good has he been thus far, and what can we project from him in the rest of the season? Coming into the season, Ryan Dempster was a very consistently average relief pitcher. Sure his ERA bounced around a lot because of the sample, but his peripheral-extrapolated ERAs from 2005-07 were 4.04, 4.19, 4.08, pretty average for a reliever. He averaged about 8 K/9, 4.3 BB/9 and about .7 HR/9. What should we have expected from him coming into the season? He's been a consistent 4.1ish ERA guy as a reliever. Going from relief pitching to starting on average adds about 3/4ths of a run to your ERA. That meshes with the statistical projection systems, so ZIPS put him at 4.98 and PECOTA at 4.89. That's three distinct methods coming up with nearly identical projections, so I'm comfortable going with that. That puts him at a little below average, as the NL ERA for starters last season was 4.64. So we know who he was before and who we expected him to be. Now we need to know who he has been so far and whether that should change our projections? Raw results: 57.1 IP, 2.35 ERA. We can dismiss the 2.35 ERA out of hand. He hasn't pitched that well. First of all, he's given up an uncommonly high total of UER, which are just as much his fault as any runs but taken off his ledger because of archaic scoring rules. So now we're at 57.1 IP, 3.45 RA. That's very good, but not quite in the elite territory of his ERA. This reflects his Support-Neutral W-L of 3.6-2.6 in 9 starts, and a team SNW% of .567 (BP hasn't updated these stats for today yet, but I did some quick calculations that should get me pretty close to the new numbers). So why has a pitcher we expected to post a 4.9ish ERA and be below average pitching like an above-average pitcher? First, and most obviously, he's been lucky. Using stats coming into today because Fangraphs hasn't updated, he's walking about the same as always, he's King a less, he's giving up HRs at a lower rate. Defense-indendent, he's been mostly the same pitcher, especially given that today's performance should bring his K-rate back up to normal. That's a good thing, considering the transition to starting should have cost him more and still might as the season goes on. If he can keep up his peripherals from his last three seasons as a reliever, he's a 4.1ish ERA pitcher, which is fine for a starter. His FIP coming into today was 3.87. But why the 3.45 RA instead of the 4.1ish ERA or the 3.87? You guys already know the answer: BABIP. He was allowing a .203 BABIP coming into today, which is simply unsustainable. How much upward should we adjust it? If we adjust it up to his career average of .309, he gives up 15 more hits, which probably means about 8 more runs. Add 8 runs to his total, and he's got a 4.71 RA, still a little better than what our projections put forth but not much. That said, I don't think a .309 BABIP would reflect how he's pitched either. He's not giving up many line drives and he's inducing a lot of ground balls. The type of BIP matters when calculating expected BABIP. Dempster's LD/GB/FB ratio (again, all stats coming into today) this season is 14.6% 56.9% 28.5%, compared with 20.4% 47.5% 32.1% career. He's giving up essentially the same FB, but he's forced 5.8% fewer LDs and 9.4% more ground balls. That's eight fewer LDs and 13 more GBs. Using the standard BABIP rule-of-thumb of LD%+.120 would give him a .266 BABIP, meaning he's allowed just 7.5 fewer hits than expected. Call it 4 extra runs to be safe. Add four runs to his actual performance this season and we come out with a 4.08 RA. I think that's even being a little harsh, because his 2+ GB/FB ratio should improve his BABIP even more. So now we know the Dempster we have. We have a Dempster who was projected to be in the 4.9 ERA range because of the transition to starting, but has instead pitched to his previous years' peripheral level plus an improved ground ball rate and thus has a defense and luck independent run avearge of about 4.1, slightly better than previous seasons. I think it's fair to say that his performance this season, adjusted for luck, has been legitimately better than projections because he hasn't lost any peripherals from his time as a reliever. So does this change our projections for him from here on out? That depends on two things: Is his newfound groundball tendency legit and can his arm hold up to starting for an entire season after relieving for so long? On the GB issue, I think there's at least some chance he can continue to induce GBs at a rate higher than expected, though perhaps not at the rate he has managed so far. As mentioned earlier in the thread, his pitching pattern and stuff has changed. He's relying on his slider much less in favor of his FB, and he's throwing all his pitches about 2 MPH slower. Fewer sliders means fewer hanging sliders, and the slightly slower fastball could mean improved movement. Both of those would lead to an improved GB rate. On whether his arm can hold up, we'll see. I'm skeptical. What does all this mean in summary? It's up to the Cubs management staff now (as much as that makes me cringe). If Dempster can keep it up for the rest of the year, great. But he needs to be monitored carefully. This good start means that his end-of-the-year totals will almost certainly be better than expected for almost the entire season, at worst around league-average for a starter. But if his GB-rate starts to slide, or if he starts to wear out and lose his peripherals, they need to be aware that he could collapse into a much below average pitcher and need to be replaced even while his season-to-date stats still look decent. He's been worth probably around a full win over expectations to this point, but that's no reason to just let him give it away in July and August if he starts to suck. I hope this all made sense.
