Who’s really the top relief pitcher in LA?

Kenley Jansen or Javy Guerra the top dog in LA?

I think Kenley Jansen—even if he records zero saves all year—will be more valuable than Javy Guerra. (I also think that Jansen, assuming he stays healthy, will be more valuable than about half the closers in the league even if he gets only five saves. But, that’s a whole other discussion.)

I don’t think this is an unreasonable conclusion, considering how much value he will generate in strikeouts, ERA, and WHIP. Just imagine what he could do if he were able to unseat Guerra as the closer!

But without him owning the closer’s role, many will ignore him. Which begs the question, is Jansen or Guerra the guy to own in LA?

While the knee-jerk reaction is always to go with the man who has the job its not that simple in this case. Fantasy owners seem to have recognized this, drafting Jansen just four spots behind of Guerra in ESPN drafts (ADP 215.4 to 219.7).

While I think the small difference in ADP has more to do with owners hoping Jansen seizes the role, he really doesn’t need the closer role to be worth the pick. All he has to do is pitch like Kenley Jansen and keep striking batters out.

It goes like this:

Regressing both Jansen and Guerra’s plate discipline characteristics and batted ball profiles, Jansen—not surprisingly—grades out as the far superior pitcher.

Actually, his numbers point to something truly remarkable.

I’ve got his final line at a 1.97 ERA and 1.02 WHIP, with a 15.01 strikeouts per nine innings (K/9) and 4.11 walks per nine innings (BB/9)—meaning he could very well be the next incarnation of Craig Kimbrel, Carlos Marmol circa 2010, or Brad Lidge c. 2004, albeit without the saves. Though he doesn’t have a clear pathway to the closer’s role, 100-strikeout setup men are a very underrated commodity in fantasy. Jansen’s overall line could be worth more than many of this year’s closers—even if he finishes the season with only a handful of saves.

Over 60 IP, I’ve got his line being worth about 0.61 points above the average reliever. Ratchet his line up to 12 saves and all of a sudden he’s worth a full point above the league average in the standings. That’s in comparison to guys like Huston Street (Steamer’s line comes in at 0.36 points) and Jose Valverde (-1.18 points by Steamer’s line) who have a full season of saves under their belt.

On the other hand, Guerra posts a more modest 3.48 ERA and 1.31 WHIP, to go along with a 7.2 K/9 and 3.7 BB/9—not far from his 2011 rates of 7.33 K/9 and 3.47 BB/9. Last year’s sparkling 2.31 ERA will likely fade, however—and with it comes the possibility of a change at closer (or so Jansen owners hope). Still, at 35 saves on the year over 65 IP, and he still comes in below average (-1.02 points).

And while we’re talking about Jansen, I think it’s worth pointing out that his control is far better than he is given credit for. Because of his high walk rates (4.36 BB/9 in 2011), he has undeservedly earned the reputation of being wild. But, that statistic is very misleading.

In actuality, he had very good control last season, posting a 53.2 percent zone rate and 59.2 percent on first strikes. The zone percentage, in particular, is very impressive for a reliever. While you wouldn’t expect a pitcher with that kind of control to walk so many batters, he does so because he generates so many swings and misses. So, instead of batters ending the at-bat by putting the ball in play (where there is no chance of a walk), those extra swings and misses keep more at-bats alive, resulting in inflated walk totals.

Bumping his O- and Z-Contact ratings up a notch (to .713 O-Contact and .858 Z-Contact—Guerra’s numbers), he all of a sudden finds himself walking 3.29 batters per nine. Going one step further, if batters chased him out of the zone at a reasonable rate (Jansen had a 25.4 O-Swing last season), he would find himself at 2.7 BB/9.

To sum up that tangent, please don’t fool yourself into thinking that Jansen is your typical high-strikeout, poor-control reliever. He’s much, much more than that!

And getting back to the main point: Be sure to take advantage of his underrated value. Jansen is just one of a handful of hurlers who doesn’t get credit for the value he contributes to a fantasy team. Though he’s the No. 2 man in LA, don’t discount him too much on draft day. You’ll regret it!

A Hardball Times Update
Goodbye for now.

What batting in a Boston (not San Diego) uniform does for your value

Lineup strength is one of the more underrated parts of a player’s fantasy value. A few things that many fantasy players do wrong is to ignore, or significantly underrate, the effect that a batter’s teammates have on his value.

Now, let’s not get carried away on that last statement. Everyone in fantasy recognizes that batting in the Yankees lineup is better than batting for the Astros. However, you get the feeling that only the keenest of owners know the true value of this switch.

Let’s take Jacoby Ellsbury for example. Out of the leadoff spot in Boston, he turned in one of the finest seasons of 2012—119 runs, 32 home runs, 105 RBI, 39 stolen bases, and a .321 average.

Absolutely extraordinary!

His regressed numbers state that he was a little unlucky in runs scored, but made up for it with slightly inflated RBI totals. That regressed line is a stunning 126 runs scored and 92 RBI. Also outstanding!

Now, for the kicker—throw him in San Diego. Keep his exact same profile intact (732 plate appearances, 32 home runs, 39 stolen bases, .321/.376/.552), but change the team around him.

How does he do?

…very well, but the lineup around him has certainly taken its toll. He cedes 18 runs and 12 RBI to finish at 108 and 80. An excellent player, no doubt, but not the same guy by any means. In fact, he cedes a full 1.7 points in the standings (similar to the difference in value between Evan Longoria and Michael Young, by many pre-draft estimates).

So, the obvious (but now quantified!) moral of the story: Think carefully about lineup strength before you draft anyone. Yes, the Boston to San Diego exchange is extreme—and I’m sure you already take lineup strength into account—but don’t forget about those five or six runs and RBI that can come from moving from, say, Toronto to Washington. Those little bits of value really add up to a lot and can make you a nightmare to play against. Add a half a point here and a quarter of point there and you’ll find yourself way ahead of the pack in September.


12 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
jsp214
11 years ago

How are you able to regress RBI/R? You have me intrigued.

I agree RBI/R can be underrated. I used to basically ignore them but you’re absolutely right. The difficulty is not putting too much weight into projected lineups since they can be shaken up so many times during the year.

One example is Drew Stubbs vs. BJ Upton. At first glance they’re strikingly similar from a fantasy standpoint. They’re both 20/40/.250 types. Upton used to hit toward the bottom of the Rays’ lineup and that was a big reason I felt he was overvalued. Last year he was moved up to 2nd and that significantly increases his value.

On the other hand, Stubbs started last season leading off but is now projected to hit 7th and that has the potential to crush his value. On the other hand, if he starts off hot or projected #2 hitter Cozart struggles, it wouldn’t be surprising for him to move up to 1st or 2nd.

Right now I have Upton about 10 outfielders ahead of Stubbs. I have a feeling that’s much too wide a gap given the fickle nature of managers’ lineup constructions.

What are your thoughts on this? Thanks!

Bobbo
11 years ago

Excellent read and I’m sure glad I have Kenley in my keeper league.  Hopefully LA wisens up and gets him in the closer role before long.

You mentioned that Jansen is just one of a handful of hurlers who don’t get enough credit from fantasy owners… do you have a quick list of a few more?  I’m thinking guys like Addison Reed and Rex Brothers will be in your list but anyone else I should target late?  Thanks!

Mike Silver
11 years ago

Totally agree on the RBI/R assessment. Also, sending a guy down to 7th in the order is pretty crushing—not only because of the lost R and RBI opportunities, but also because they see significantly fewer plate appearances.

A quick rule of thumb (without any other math in there) is to assume 18 plate appearances lost for each spot down in the lineup per 162 games played. To do this quick computation, assume everyone on the team has the same number of PAs until the last out is made. So, if you go through the lineup 4 full times, then the game ends on the 5th time through on the 3rd batter—only the 1st and 2nd hitter get 5 PAs and everyone else gets 4. If you assume that the batter who makes the final out is pretty random, you get the following: 162 games / 9 lineup spots = 18 PAs. Therefore, you lose 18 PAs per time through the lineup. A more accurate equation would take in OBP, but this one is reasonably accurate and probably isn’t more than 5 PAs off per 162 games.

So, start at 756 PA for leadoff, then 738 for 2nd and 720 for 3rd. By the time you get down to 7th, you’ve lost 108 PAs (down to 648 PAs). That’s an enormous difference. Now, Stubbs and Upton have a huge divide since Upton has 108 more PAs to hit HRs and steal bags. Then, factor in the increased R/PA at that position (though there is reduced RBI/PA at leadoff) and you’ve got a huge difference.

I agree on the fickle managers as well. However, with certain guys, you can estimate playing time over the year and at least get a rough estimate.

As for the regression equation itself, below is the one I used for leadoff hitters. Note that it is regressed to R/PA and RBI/PA, so you have to multiply up by their estimated PAs. Also, this particular equation has been regressed for leadoff hitters only, so it will give screwy values if you try to plug in for a different lineup spot:

Leadoff R/PA: .2029907*AVG + .2572099*ISOOBP + .1495043*ISOSLG + .1476099*SB/PA + -.0028599*(AL=1, NL=0) + .3498121*TeamAVG + .1611112*TeamISOSLG +-.0738139

Leadoff RBI/PA: .040348*AVG + -.1358064*ISOOBP + .0460996 * SLGRatio + .0050491 * (AL=1, NL=0) + .7566887*HR/PA + .3375898*TeamAVG + .2439369*TeamISOOBP + -.0253133*TeamSLGRatio + -.0673156

Another note:

-ISOOBP is isolated OBP (like isolated SLG%, which is just OBP – AVG = ISOOBP)

-SLGRatio is the ratio of slugging percentage to batting average. I find this to be more correctly representative of a player’s power than ISOSLG, since ISOSLG attempts to measure a player’s power but does not stay constant by batting average. Here’s an example:

Think of two batters who hit only doubles. You would argue that they have equal power because they hit a double every time they come to bat.

Batter A has a batting average of .250
Batter B has a batting average of .300

Batter A, therefore, has a SLG% of .500
Batter B, therefore, has a SLG% of .600

Batter A has an ISOSLG of .250
Batter B has an ISOSLG of .300

Using the typical interpretation of ISOSLG as a measure of a player’s power, you would conclude that Batter B has better power than Batter A. However, if they both only hit doubles, you can make the argument that they are equal.

So, if we’re discussing which player has better power, you would conclude Batter B with ISOSLG—when that has more to do with better batting average than raw power.

jsp2014
11 years ago

wow, awesome Mike! that is going to be incredibly helpful for me. I can’t thank you enough for your detailed response.

Bobbo
11 years ago

That analysis makes me cringe at my decision to keep Jason Heyward (just kidding… or am I?)  Get him out of the 7th spot Fredi!

Mike Silver
11 years ago

No problem guys!

DShea
11 years ago

I think a key question on lineup and team change is whether projection systems use that info.  Most do adjust for team changes, right? If a system uses batting order in projections its wrong that can be big

chongo
11 years ago

A 4.11 walk rate is a 4.11 walk rate- not good unless some were IBB which I doubt in Jansens’s case.  Everybody is so in love with his strikeout rate, but it is likely to come down;  sure he could do a Kimbrel 2011, but it is likely that even Kimbrel can’t do a Kimbrel 2011.

Mike Silver
11 years ago

Can’t disagree that a 4.11 walk rate is bad. The point I want to make is that he doesn’t have poor control, which a lot of analysts claim he does.

Even so, his exceptional strikeout numbers make it OK to have a high walk rate. If no one can put the ball in play, the only way you can score is to walk 4 guys in one inning.

Kimbrel, the old Marmol, and Jansen are classic examples of people focusing on the bad instead of the (lots and lots of) good.

Jay Dub
11 years ago

Wow.  Awesome article, thanks!  Curious about one thing – what’s the math behind using the O- and Z-contact and O-Swing to project a BB/9 ?

“Bumping his O- and Z-Contact ratings up a notch (to .713 O-Contact and .858 Z-Contact—Guerra’s numbers), he all of a sudden finds himself walking 3.29 batters per nine. Going one step further, if batters chased him out of the zone at a reasonable rate (Jansen had a 25.4 O-Swing last season), he would find himself at 2.7 BB/9. ”

Mike Silver
11 years ago

Hey Jay Dub.

I wish I could give you a good answer to your question, but the K/9 and BB/9 come from a really complication regression formula that doesn’t fit in the comment box. There’s a lot of inputs, so it’s tough to describe here.

I’m guessing you can pick up on the idea of it though: if batters make more contact on pitches outside the zone, you walk less since that’s a pitch that was turned into a ball in play—as opposed to keeping the at-bat alive with a strike/ball. If you bump up the Z-Contact, you get the same effect.

If it helps, here are the coefficients for O- and Z-Contact for K/PA and BB/PA (which the formula is based off of, then you have to do a few more adjustments to get it into K/9 and BB/9—another reason why its too long for a comment section. However, if you estimate batters faced and innings pitched, you can do some back of the envelope math with different players):

Z-Contact:

K/9:  -0.81676
BB/9: -0.16316

O-Contact:

K/9:  -0.25272
BB/9: (Not a statistically significant variable)

I guess I could have left out O-Contact in the explanation above then, seeing as it wasn’t statistically significant in the model anyway! See, even I can get lost in my own model!

Ben Pritchett
11 years ago

Excellent work, Mike. I’m left wanting more.