Notice: Trying to get property 'display_name' of non-object in /var/www/html/wp-content/plugins/wordpress-seo/src/generators/schema/article.php on line 52
keyboard_arrow_uptop

New Year’s Resolutions, diet plans, and closers all share one thing in common–a significant fail rate. Research by Ron Shandler of BasebalHQ.com shows that over the past twelve seasons, the lowest fail rate of closers in a fantasy baseball season happened in 1999 when just 22 percent of drafted closers lost their job in a season. Since 1999, that rate has been anywhere from 22 percent to 59 percent as closers tease fantasy owners more than the cute girl in middle school passing love notes requesting a check box to be filled in.

If you have been playing this game long enough, you have invariably heard every fantasy tip known to mankind when it comes to closers. Some of my favorites:

  • Saves are from the devil, Bobby Boucher!
  • Punt saves, they’re a sucker bet and you do not need them to win
  • Saves are one-tenth of your scoring–they’re equally as important as any other category and you should not slight them
  • Just draft three stiffs in the final three rounds or spend $1 on three different relievers. Saves come from everywhere

The good news for the role is that it had a 75 percent success rate last season as most of the failures were limited to the Chad Qualls meltdown in Arizona, Trevor Hoffman aging ten years before our eyes in Milwaukee, and Mike Gonzalez breaking down quickly in Baltimore. For the most part, closers did their job well and piled up saves but some closers such as Francisco Cordero, David Aardsma, and Octavio Dotel did so while dragging down owners’ ratios. Coming off a year in which closers finally had a low fail rate, fantasy owners may be a little less weary about treading back into the dangerous waters where high picks and high dollars are spent on closers.

In a current “expert” mock draft that I am in that includes names like Cory Schwartz and Mike Siano of MLB.com as well as the esteemed Joe Sheehan, the first closer went off the board early in the sixth round in the form of Carlos Marmol. Five more went before the end of the seventh round. For comparison’s sake, my first closer pick was the first pick of the twelfth round when I selected Francisco Rodriguez 133rd overall. I have had a mixed history with closers as I have both won leagues running away with saves and also won a 10 team AL league with just 13 saves on the entire season as early injuries led me to stocking up on starting pitching rather than be bent over a barrel by trade partners trading from a position of strength.   

No matter which path you decide to take with closers, you must have a plan of attack for them on draft day. My personal preference is to find the closers that safely meet the skill set of what I consider a desirable closer. I want closers that keep people off the bases by striking them out and limiting free passes and that throw a lot of strikes.  I also want closers that keep the ball on the ground and limit their home runs which are the quickest way to blow a save. Lastly, I want guys who do not have a demonstrable weakness against righties or lefties as those types of pitchers tend to get put back into limited roles when they struggle.  Each draft season, I put together a list of the top relievers available in the draft pool and put their statistics into a matrix that contains the following categories:

  1. K/9
  2. K/BB
  3. GB%
  4. HR/9
  5. Difference between their best and worst split in batting average against righties and lefties
  6. STR% – the percentage of pitches they throw for strikes
  7. K% – the percentage of the plate appearances the opposing batter strikes out in

The table below contains those metrics of the 40 different potential closers that Mike Petriello ranked earlier this week.  The first four categories are the projected totals from our own Player Forecast Manager while the final three are the career average for each pitcher. While it is easy to just work down Mike’s rankings one at a time, sometimes rankings do not tell the full story. The matrix below has been color-coded for both strong indicators as well as cautionary indicators for each one of the seven specified skills mentioned above–green is good, red is bad, and those left blank are somewhere in between.

 

You can also access the spreadsheet through this link.

Brian Wilson gets top billing in the rankings and it is for more than just his awesome facial hair. The chart shows that 25 percent of the batters he has faced in his career have walked back to the dugout mumbling something under their breath but what it does not show is that his K rate has risen each of the past five seasons: from 16 percent back in 2006 up to 30 percent in 2010. Even with that strong performance, there is an argument for passing on Wilson and taking Joakim Soria, as he performs extremely well in nearly every one of the categories. Note how arguably the best closer in baseball is ranked just fifth for fantasy purposes but shows no areas of concern in his career rankings. However, Rivera’s K rate is down three straight seasons and it dropped a full eight percentage points from 2009 to 2010. That is not something that only older pitchers go through, as Jonathan Papelbon has seen his K rate drop for four consecutive seasons from an unworldly 38 percent down to just a strong 27 percent.

I prefer to use this matrix style of analysis with color coding because it quickly allows me to review the strengths and weaknesses of a reliever as I evaluate him mid-draft. If I am late in a draft and looking for another closer, I am more likely to jump in on Leo Nunez because he shows one of the stronger skill sets of the lower ranked closers over someone like Brandon Lyon or Ryan Franklin who are ranked higher but present several areas of caution.  I also know that Nunez’s K rate has increased each of the past three seasons.  Additionally, the picture helps you make up your mind when staring down two veteran closers such as Brad Lidge and Jose Valverde who are closely ranked but one comes with a lot more statistical risk than the other one does.

This type of matrix also helps when you are considering two different relievers from the same team. Find David Aardsma and Brandon League on the matrix above–does either one stand out above the other one? That is a situation that could go either way once (or if) Aardsma is fully recovered from his injury or Eric Wedge can simply let Aardsma keep his job since League is not demonstrably better statistically than Aardsma. However, compare the data for Jonny Venters and Craig Kimbrel with the caveat that Kimbrel’s data was pulled from his minor league work (courtesy of statcorner.com).  Note the large difference in dominance that the PFM projects for the two relievers as well as the difference in K/BB ratio. Kimbrel has the risk of large splits between righties and lefties, but the quick skills picture shows him as the projected better pitcher.

Auction bids happen quickly and closer runs can be over before you even realize they started. Having a quick and dirty way of scanning the available talent pool before making a decision on whether to jump in or sit back allows you take the time you would have spent comparing lines on a spreadsheet and focus on your next two or three roster moves in the draft as savvy chess players do in a match. I strongly suggest you print out this color grid and use the white area out to the side to include your own player notes that I mentioned earlier with regards to Wilson, Rivera, and Papelbon. Or, you can add that Joe Nathan’s K rate was in a three year rise before his injury, that Matt Thornton’s K rate has also increased each of the past five seasons, or that Jose Valverde’s has declined each of the past five seasons. Closers are a risky business and the more information you have at your side on draft day regarding their past performance, the better off you are with the standard caveat that past performance does not guarantee future returns.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
mhmosher
2/25
wow..for a fantasy junkie, this is great. I am resubscribing partly based on Jason's work here.
mpet29
2/25
Jason, this is really great. Wish I'd thought of this myself... and mostly happy that my rankings mostly correspond with your research. Very cool.
moonlightj
2/25
Wow, thanks guys.

I've done this kind of thing for a few years in my own Excel sheets for prospect tracking. I'd conditional format each column on the pitcher tab and hitting tab so I could get a quick glance at warning spots as I inputted the statistics and then do some asking around on scouting reports or in the case of Florida State League guys, go check them out.

I found it was a nice switchover to fantasy drafts and if enough people find this format popular, I'll work on the other ranked positions that Mike and Marc have put up recently.
moonlightj
2/25
Marc included the link to the data if you want to pull down your own copy. My parameters for "good" and "worrisome" might be different from yours as I only greenlighted K/9 rates of at least 9.0 for closers but 8.1 (Nunez) might be good for you as well. I set a 7.0 K/9 baseline for my SP and a 9.0 K/9 for my closers.
moonkyu
2/25
Yes, would love to see quick reference guide for position players and starters. This would provide an easy way to differentiate between players within same tiers. Super helpful.
granbergt
2/26
Agreed
alscherer
2/28
I absolutely agree! This is excellent.
adrock
2/25
This is excellent writing, Jason. The analysis is very impressive, and should be quite useful to anyone playing fantasy.

Will there be similar analysis of middle relievers, for those of us in deeper leagues?
moonlightj
2/25
Middle relievers are doable. I'll pull the top 40 from the PFM and put them into a piece.
Stingers56
2/25
Excellent, I play in a league where middle relievers are relevant.
coryschwartz
2/25
Jason, great stuff and very well-researched as always, and thanks for the name-check!

However, the main disagreement I have with the "don't pay for saves" strategy is the logical fallacy that the failure rate among closers is linear; that is, all 30 closers have an equal risk of failure. Intuitively we would expect this to be untrue and factually it can be demonstrated to be untrue. I demonstrated this back in '09 (http://bit.ly/apW4xj) and while your mileage may vary from year to year, I think the conclusion is plain, that the more expensive closers are so in part because they are more reliable and generate a better ROI (in the aggregate) than bargain closers.

For every Joe Nathan or Jonathan Broxton from last year, you will find several bottom-third closers who failed, not to mention the fact that the upper tier guys generate more K's and better ratios than those at the bottom.

Bottom line is, in the aggregate, you actually do get what you pay for with closers. If you want to bargain-hunt, you may indeed find a bargain, but you have a greater chance of doing more damage than good than if you had simply invested in a more proven commodity.
moonlightj
2/25
Good points, Cory. There is no problem paying for consistency as the fail rate in the top 10 projected $$ slots is going to be much lower on average than the bottom 10 guys.

For my own comfort level, I'd rather chase saves throughout the year and draft closers after round 10 than have to chase homers or steals because I passed up good sources of those by overdrafting a closer in round five.

I could take K-Rod, Leo Nunez, and Joel Hanrahan in a mixed league and be very happy with my end results as all three would have been taken from the 12th round on back. It would give me enough saves to compete in the category while really helping my strikeout totals. Meanwhile, I've seen other drafts where people have taken two closers before the 12th round which I am just not comfortable doing due to past experience.
coryschwartz
2/26
No doubt, past experience is also critical. My two best teams in NFBC were in '06 and '09, when I had dominant bullpens to support loaded offenses and bargain-basement SP rotations. There are only 30 closers to choose from at any given point, but ~150 starters, so I'd rather sift that pile instead.

Ultimately though it's a question of risk tolerance and reading the depth of the player pool. Looking at all of the uncertainty in so many bullpens this year certainly creates buying opportunities, but for those who bet the wrong horse, it could be a long season of chasing saves...
Stingers56
2/25
There's no question you get what you pay for, generally speaking. If you follow the high closer route, you're invested in them and hopefully you get what you paid for. And there's a good chance that will happen, barring the unforeseen.

The reason I like the cheap closer strategy is that it positions you to be on the look out for the next emerging closer to be. Sometimes those guys turn out to be really good. That's not a bad return on investment. High risk, high reward for one stat in a 5x5 format.
mhmosher
2/25
In all honestly, no bull, Jason is putting out some of the most valuable fantasy info on the internet. This is great.
SFiercex4
2/25
Just going to reiterate how awesome this piece was and in particular how great that table was. Thanks Jason!
TheManster
2/27
Neftali Feliz had a 3.94 K/BB. The 2.2 was his BB rate.
moonlightj
2/27
Thanks for that catch - it has been corrected
cavebird1
2/27
Jason, great article, interesting stuff. However, I think there is one big caveat you missed in the Kimbrel/Venters comparison. The PFM (PECOTA) projections for Venters appear to be based on both his season last year and his minor league numbers, which makes sense because he was a rookie last year. However, in 2008 and 2009, Venters was used almost exclusively as a starter in the high minors, and his numbers, especially his K numbers, are not surprisingly a lot lower. So, I wouldn't put much stock in the K% and K/BB numbers that the PFM projects for Venters.
baserip4
2/27
"The good news for the role is that it had a 75 percent success rate last season as most of the failures were limited to the Chad Qualls meltdown in Arizona, Trevor Hoffman aging ten years before our eyes in Milwaukee, and Mike Gonzalez breaking down quickly in Baltimore." Plus Frank Francisco.

And that is small consolation to the owner (me) that had Qualls, Francisco, and Gonzalez.
moonlightj
2/28
I was taking Qualls as a 2nd closer in a lot of leagues last year - I'm right there with you.
Robotey
2/28
2nd closer! Consider yourself lucky! In my 11 team NL-only I targeted him at $17 as my primary closer. Five picks into the draft someone throws him out there, so I follow my plan...bid...$17, not going any higher I tell myself. and I don't. And neither does anyone else, so he's mine at $17. All according to plan and on budget: until he goes and stinks up all of April. And May. And June and you know the rest. Perhaps you ought to write about tapping our inner Kenny Rogers (not the pitcher) and knowing when cut losses and pull the plug on a closer when's got nothing but WHIPS that look like ERA's waiting for you. This year I'll stick with keeper Brad Lidge at $11 and pray.
sharbrough
3/01
I notice that no pitcher was "green" in the three categories of GB%, STR% and K% (an all-green "state" for those three measures). I wondered what such a guy looked like.

But there are a few guys who are green in two of the three categories, and white in the third (a group of 3 states for the same measures). As I looked at those players, they seemed like good bets to limit risk, if the draft position was reasonable.

Have you looked backward and compared the "state" of the indicators to the $ earned in prior seasons? In short, is there some predictive value to these states?
moonlightj
3/01
I have not yet taken that look back at the end of the year against this but will put it on my to-do list for this season.
hessshaun
3/01
I want to say that when I originally read this, there was no chart? I love the chart and my PFM looks very similar to this, but per my stat categories. I did not think to isolate by position to kind of dig for gold, which is really easy to do in one fell swoop of selecting categories. Needless to say, I love the idea because the color just makes the PFM look all the more interesting and it's nothing more than sorting and shading. Hopefully I can employ this to find some random gems at other positions. Also, the other thing I do is mark people yellow, for injury risks, and or pitchers who have any sort of discomfort at the end of last year or during Spring Training. Thank you MLB_ID.

Great idea, one quick question and something I always struggle with. Where did you come up with the arbitrary number to make the demarcation per category? How did you determine that? I normally sort the entire PFM dump and always struggle with where my green and red start.
moonlightj
3/02
A mix of personal preference and league averages. For K/9, I'll dip down as far as 6.0 for starters but I want closers that miss bats at strong rates. GB% is pretty standard as the demarcation between groundball and flyball pitchers and the rest are based off league averages where I green light those 2% points above the league average and red light those 2% below it.