Tagged: xFIP

June 25, 2010

Should You Sell High on Ubaldo Jimenez?

By Eriq Gardner //

If the season ended today, Ubaldo Jimenez would win the Cy Young award easily on the strength of a 13-1 record, a 1.60 ERA, a 1.04 WHIP, and 95 strikeouts.

He’s been the top player in fantasy baseball this year. Yet there’s a creeping suspicion that he may be a little overvalued.

Let’s examine the evidence:

He’s averaging about 7.99 strikeouts per 9 innings. That ranks him 26th among starters with at least 70 IP this season. He’s behind Colby Lewis, Ricky Romero, and Gavin Floyd, to name three pitchers with much less star power.
He’s averaging 3.03 walks per 9 innings. That ranks him 54th among starters with at least 70 IP. Among the pitchers walking batters at a lower rate are Brian Bannister, Kyle Kendrick, and once again, Gavin Floyd.
He’s averaging 0.34 home runs per 9 innings.That ranks him 6th among qualified starters. However, only 4.4% of his fly balls are going for home runs. His career rate is 7.6%, and he plays half his games at Coors Field. That’s very likely to regress.
His strand rate is 87.8%, meaning 87.8% of the runners he puts on base don’t score. That’s largely a function of fortuitous timing (pitching abnormally and probably unsustainably well with men on base), as well as unusually strong bullpen bullpen. That’s also the luckiest rate in the entire major leagues.

Add it together, and Jimenez’s xFIP (a measure of ERA independent of various luck factors) is 3.64 — more than two whole runs higher than his real ERA. For the rest of the season, Bloomberg Sports projects Jimenez to produce 9 wins, a 3.39 ERA, a 1.21 WHIP, and 112 strikeouts.

People toss around the phrase, “regression,” but what does it really mean? If Jimenez’s peripheral stats indicate a great deal of luck, should owners sell him now?

It depends.

First, nobody should succumb to the “gambler’s fallacy,” or an expectation that Jimenez’s great luck thus far will be evened out by unusually poor luck upcoming. Regression only means that based on peripheral stats, one should expect an ERA at about 3.3-3.6 from this point to the end of the season — and not that he’ll end up at that result at season’s close. An ERA at this level is still phenomenal. Bloomberg Sports projects him as the third-best pitcher in baseball from this point forward.

Second, deciding whether to sell him “high” depends on context.

If you own Jimenez, do you still need great pitching? If so, it’s going to be hard to do better than what he’ll give you.

What’s the trade offer? The presumption that he’s a sell-high candidate because an owner can trade Jimenez for a comparable pitcher like Josh Johnson, plus get something extra, although in reality, it’s not a slam dunk that a Josh Johnson owner would make this deal.

That said, perhaps there’s still an opportunity to trade Jimenez for more than what he’s really worth, to an owner who needs a pitcher and is willing to exchange a batter. In the preseason, hardly anyone gave any thought whatsoever to drafting a pitcher with the #1 overall pick. But now that Jimenez sits atop fantasy player raters as the best in baseball, it’s perhaps conceivable that a pitching-starved team trades Hanley Ramirez, Albert Pujols, Ryan Braun, or another top-notch batter for Jimenez.

Lastly, owners shouldn’t be entirely sure that selling Jimenez now is the smart course of action, because there’s no guarantee that the Rockies’ ace really is lucky.

Sure, all the best sabermetric stats at the moment indicate he’s doing better than he really deserves, there’s always the chance that a player defies the odds for longer than a few games or a couple of months; they might even keep the good times going all season long.

For example, if I was making the case that Jimenez really isn’t quite as lucky as he seems, I’d point to the fact that his command improves when batters get on base. (His walk rate goes from 4.03 with the bases empty to 1.84 when he’s pitching from the stretch.) In turn, his xFIP in these situations drops nearly an entire run. This could partly explain why his strand rate is so tremendous. With men on base or in high leverage situations, Jimenez simply performs better. That’s probably not sustainable over the long haul, but it is possible.

Plus, peripheral stats do a good job of telling us general trends among the general population of players, but can often miss the mark when pinpointing individuals. What’s the expected batting average on balls hit in play on a pitcher whose average pitch velocity is at least 96 miles per hour? Hard to say, because there’s only been one pitcher this past decade whose reached those heights. Jimenez, of course.

So if you’re thinking about selling high on Jimenez, remember to consider context when making or fielding offers. Getting a big haul would be great. But don’t trade him just for the sake of trading him.

For more on Ubaldo Jimenez and other pitching aces, check out Bloomberg Sports’ Fantasy Tools.

March 16, 2010

Dan Haren’s Fantasy Value: Don’t Fear the Splits

by Eno Sarris

Sometimes a player just can’t get any respect, despite consistent excellence. That might be the case with Arizona ace Dan Haren.

The good news is that over the last three years, he’s posted ERAs of 3.33 or lower, WHIPs of
1.21 or lower, microscopic walk rates of 1.8 per nine innings or lower, and a strikeout rate above 8 K/9 IP. FanGraphs’ Expected Fielding Independent Pitching stat (xFIP), which runs along a similar scale to ERA and isolates factors a pitcher can best control such as home run rate, walk rate and strikeout rate and adjusts for park effects, defense and other factors, shows Haren ranking 4th in MLB in 2009 (3.08) and 4th in 2008 (3.21). Consistency is part of Haren’s oeuvre.

Take a look at Bloomberg Sports’ Fantasy Baseball Kit and several other numbers jump out at you. Haren was better than 75% of starting pitchers in strikeouts, and also sported an elite WHIP. Toiling for the 70-92 Diamondbacks kept Haren from amassing an impressive number of wins (he settled for a still solid 14), but there’s room for hope there as well. The D-Backs’ 720 runs scored last year were average for the National League (8th out of 16) and with a few steps forward from youngsters Justin Upton and Stephen Drew, the team could add to that total – especially in a hitter-friendly park. In the end, though, those factors don’t matter all that much. If Haren nets another season with the 16th-ranked ERA, first-ranked WHIP, and sixth-ranked strikeout total in baseball, you’ll be happy.

Still, much has been made about Haren’s first half vs. second half splits; the Bloomberg Tool again shows us graphically what the naysayers are saying. You can see that Haren’s ERA rose as the season wore on in 2009. This is not a new development. Take a look at his career pre- and post-All Star splits: In 651 pre-All Star innings, Haren has a 3.08 ERA, 1.06 WHIP and a 7.45
K/9. In 575 post-All Star innings, he has 4.21 ERA, 1.32 WHIP and a
7.76 K/9. Those numbers include a mediocre career 4.58 ERA and 1.31 WHIP in August.

But that doesn’t mean you should draft Haren and trade him at the All Star break, or worse, pass on him entirely. The pre- and post-All-Star K rates were an early clue. But if you take ERA and WHIP out of the story and focus on underlying statistics, Haren is not really any different after the break, or even in his supposedly worst month of August.

To strip luck and other factors out of the equation, let’s return to Fielding Independent Pitching, the brainchild of baseball researcher Tom Tango. The expected version of FIP, xFIP, normalizes for home run rates given a pitcher’s ballpark and league situation. Listed on FanGraphs.com are Haren’s month-by-month xFIPs for his career: 3.82, 3.72, 3.73, 3.54, 3.37, 3.64. That’s right, over his career, Haren actually has his lowest monthly xFIP in August.

So why the ugly ‘regular’ stats? Since xFIP normalizes home run rate to a pitcher’s career average, it removes the effect of Haren’s biggest August problem – his 1.58 home runs per nine innings. His career home run rate is 1.03 home runs per 9 IP. You might argue that Haren tires late in the season and is more prone to the big fly, but there’s a fly in that ointment. Haren’s career home runs per 9 IP in September and October? 1.01. It simply doesn’t make any sense to say that Haren is tired in August but fine in September. For the more statistically inclined, there’s even a numbers-based argument against the importance of season splits like Haren’s here. Basically, they don’t exist.

But even those that don’t want to read about “r” values and other advanced statistical measures can see that it doesn’t make much sense to worry about one month. That’s doubly true when the underlying numbers don’t support the case for that one month being fundamentally worse than others.

Bloomberg Sports rates Haren as an elite starting pitcher, giving him a B-Rank of 29. That’s 5th among all starting pitchers, trailing only Tim Lincecum, Roy Halladay, CC Sabathia and Zack Greinke. Draft Haren with confidence, and don’t trade him unless you get a top-tier offer. Not even at the All Star break.

For more on why Dan Haren is good, and the horde of other starting pitchers that aren’t as good, check out Bloomberg Sports’ fantasy kit.