Predicting Holds

Adding Holds as a category to any fantasy league opens up a new pool of pitchers to the mix. No longer are teams focused just on just starters and closers, now middle relievers have more value. Holds are an impossible category to predict, but today I will give some ideas about guesstimating Holds and some pitchers who may accumulate them in 2014.

If a reliever comes into a game to protect a lead, gets at least one out and leaves without giving up that lead, he gets a hold.

MLB Miscellany: Rules, regulations and statistics

Two things must exist for a pitcher to a get a Hold, opportunity and talent. Starting with opportunity: Holds are just like Saves, the manager must trust the pitcher enough to put them in a Hold situation. Two types of pitchers are given Hold opportunities, stud setup men and LOOGYs. A setup man will pitch the seventh and/or eighth since their peripheral data is similar to a closer. They will throw one inning and be done. LOOGY’s will be used to get one or two tough left-handed batters out and be done.

Like closers, who is and isn’t a setup man changes all the time. During the season, the best source for closer data is our own Bullpen Report. The rankings are update daily with changes in bullpen depth and provide the newest news on any changes. Before the report is available daily, you can also find depth charts on FanGraphs.

The other key to Holds, besides opportunity, is talent. Does the pitcher have the ability to keep the lead? To get an idea of the quality of pitcher which usually gets holds, I looked at the first 30 and the next 30 Hold accumulators over the past three seasons. Here are the results:

Hold Leaders (1 to 30)

Season HLD K/9 BB/9 IP/G IP ERA SO WHIP W
2013 24.6 8.5 2.9 0.94 66 2.79 63 1.15 4.7
2012 25.6 9.2 3.0 0.89 63 3.04 67 1.19 3.2
2011 23.9 9.1 3.0 0.93 63 2.75 64 1.10 3.4
Average 24.7 8.9 3.0 0.92 64 2.86 64 1.14 3.8

Hold Leaders (31 to 60)

Season HLD K/9 BB/9 IP/G IP ERA SO WHIP W
2013 15.8 8.7 3.3 0.87 53 3.39 51 1.26 3.2
2012 17.1 8.5 3.2 0.86 56 3.12 52 1.20 3.1
2011 14.9 8.0 3.4 0.94 61 2.97 56 1.24 4.1
Average 15.9 8.4 3.3 0.89 57 3.16 53 1.23 3.5

The Holds leaders group have an average K/9 of 3.0 and a sub 3.00 ERA. The second group is not as talented with a 2.5 K/9 and 3.16 ERA. Each list contains some true setup men and LOOGYs. Splitting the data into each pitcher type, here are the final numbers (22% are LOOGYs).

Pitcher type HLD K/9 BB/9 IP/G IP ERA SO WHIP W
LOOGYs 18.3 8.8 3.4 0.71 48.6 3.22 45 1.26 2.3
1 inning Setup Men 20.5 8.6 3.0 0.95 64.4 2.95 61 1.18 4.0

LOOGYs produce just a bit less than true setup men. Their main downgrade is the lower number of innings and therefore fewer strikeouts and Wins.

Looking at the above tables, you can get a good idea of the number of possible Holds a pitcher may accumulate. The top setup men will have around 25 holds. The next tier down 16.

To help determine the relievers with the talent to be a top setup man, I ran a query on the relievers projection by Steamer and/or Oliver. The relievers selected have a projected K/9 > 8.0, a BB/9 < 3.5 and the 90% of their games will be as a reliever. In all, 99 pitchers made the cut and the full list is available in this spreadsheet (2013 team is listed and most of the closers are included).

Some setup men are projected, according to ERA, to perform better than the team’s closer. Number eight to ten, Mark Melancon (Pirates), Joaquin Benoit (Padres-now), and Sean Doolittle (A’s), fall into this group.

Holds may not be the easiest category to predict going for pitcher. Data from previous seasons is basically useless. To find pitchers who will accumulate Holds, look for pitchers who will be give the opportunity to setup and are good. Once your locate these pitchers, turn to historic Holds numbers for more advice. With a little effort, you can really take advantage of the lack of information on Holds and make informed decisions.




Print This Post

Jeff writes for FanGraphs, The Hardball Times and Royals Review, as well as his own website, Baseball Heat Maps with his brother Darrell. In tandem with Bill Petti, he won the 2013 SABR Analytics Research Award for Contemporary Analysis. Follow him on Twitter @jeffwzimmerman.


24 Responses to “Predicting Holds”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Jeff says:

    Terrific research here, but a couple of wording issues in the 2nd paragraph.

    Vote -1 Vote +1

  2. Kam says:

    I can’t seem to access the spreadsheet.

    Vote -1 Vote +1

  3. Cory says:

    Can I get clarification about the hold rule: it seems that Holds are only awarded when the lead is 1-3 runs (similar to a save situation). At least, I have encountered that in fantasy, but the rule doesn’t appear to make this a requirement. Can an official scorer ignore the MLB rule?

    Vote -1 Vote +1

    • Hendu for Kutch says:

      It’s not that it has to be a 1-3 run lead, rather they must come in in a save situation. So yes, any lead of 3 runs or less, but also applicable would be leads of more than 3 where the tying run is in the on deck circle.

      So, for example, if your reliever came in with a 5-0 lead and the bases loaded, he’d be eligible for a hold as long as he recorded an out and the lead wasn’t gone when he left the game.

      This can lead to the occasional oddity. In the previous example, the reliever could give up a grand slam, strike a guy out, then walk the bases loaded before being relieved himself. In that case, he’d qualify for a hold despite the horrendous performance because the save situation + recorded an out + left with lead criteria are all met.

      Vote -1 Vote +1

  4. Rob says:

    I like the analysis on holds. It is a wide open topic with little research, Some other insights that would be good to see is a look at manager styles that use there pens a bit more. The idea that in order to get a hold a team must be winning, so looking at winning teams. Teams with stronger closers to fend off the holds/strong set up men from supplanting them. You mentioned holds as a balance of opportunity and stats. Stats are the easy part research, the opportunity is more difficult and hence, more valuable research. Thanks!

    Vote -1 Vote +1

  5. Can't get in says:

    Yeah, it says i need access to the spreadsheet. You need to open up permissions on Google Docs. I’d love to see what you came up with here!

    Vote -1 Vote +1

  6. KB says:

    Still requires permission to access spreadsheet.

    Vote -1 Vote +1

  7. Cole says:

    Love the research, is there a way to play Holds based on Manager style? Are there some managers or maybe even just organizations that just produce more holds or a holds just a by product of wins? If we think about it like saves, even closers on bad teams can rack up 30-40/year if they don’t lose their job or get traded.

    Vote -1 Vote +1

    • Jeff Zimmerman says:

      I never usually worry about team quality (maybe use it as a tie breaker). I would stick with a good pitcher on a bad team who won’t kill your rate stats, give good strikeout numbers but may get 3-4 less holds.

      Vote -1 Vote +1

      • exxrox says:

        I would argue that bad teams and the managers who run them do play a nice part. I wasted a lot of time waiting for the fantastic Steve Delebar and Brett Cecil to accrue my holds last season but they combined for next to nothing compared to Darren Oliver.

        Vote -1 Vote +1

  8. sportsczar says:

    Does Brian Wilson’s lack of recent data keep him off the list, or were the numbers just not there? I suspect he’ll have plenty of opps. for holds making $10M this year. I love Paco Rodriguez and am glad to see his numbers put him in the top third of your list. But I can’t help but think that Wilson will also have a ton of holds this year.

    Vote -1 Vote +1

  9. sportsczar says:

    Sorry, I’m an idiot. I glossed right over the actual numbers you used in the article. I could just look at Wilson’s projections on my own. I see that both Oliver and Steamer have him over 3.5. ZiPS has him just below, at 3.37, but that’s not exactly stellar. However, ZiPS has his K/9 over 9, so it seems pretty favorable for Wilson, while Steamer and Oliver are a bit pessimistic.

    Vote -1 Vote +1

  10. ettin says:

    Holds are used in both of my dynasty leagues and I exploit this every year. Usually in the draft or through trade I will keep at least one closer but I then focus on high quality set up men (particularly the ones Jeff mentioned that are better than the actual closers in terms of peripherals and K/9) through trade or in the draft.

    It saves me draft dollars and helps me to win a category that is exactly the same as the Saves category. I’ll pick up a closer or two later on in the season off the waiver wire or through trade if I need to close the gap in the Saves area.

    Really it just comes down to additional dollars saved for other draft targets. Most other owners focus on the Saves category too much and pay the higher dollars for it.

    Vote -1 Vote +1

  11. ms says:

    I would recommend reading the daily box scores and looking to see usage. Most managers have at least some kind of bullpen hierarchy. The later in the game the hold occurs, the more predictive the future hold probability. Also, holds involve playing the hot hand a lot…you can also look at the last 10 outings a reliever has had to get an idea of usage.

    Some teams seem especially slavish to a bullpen hierarchy. The Braves/Freddi will have a designated setup guy…Venters and O’Flaherty were absolute holds machines the last few years. Luke Gregorson set the SS record for holds (40)a few years ago (2010, I believe) and the Padres/BB seem like they get a lot.

    My top hold guys in the NL for 2014 would probably be Wilson, Clippard, Melancon, Benoit,C.Martinez (if he doesn’t start) and whoever looks like they are healthy out of Walden, Venters and O’Flaherty. Maybe also Sean Marshall if healthy or Sam LeCure.

    I play in a 12 team NL only league that’s had holds the last 5 years or so. Holds are a terrible real life stat obviously but they add a lot to FBB.

    Vote -1 Vote +1

  12. Jose says:

    How to win holds:

    1. Draft Clippard or Melancon

    2. If they displease you during the season, go to the waiver wire and find the guy with the most holds. Pick him up.

    I play in a league with holds and people are still lazy about drafting or picking up middle relievers. There should also be several holds candidates per team, but only one closer. Plenty of waiver wire fodder to go around.

    Seriously though, there are a couple of other factors I might look at:
    Innings Pitched: Guys that are used more often or are out there longer get more holds.

    Who is the closer? If the closer is shaky and your holds guy is next in line, that’s not great. Most fantasy players rejoice when their guy becomes a closer, but in a holds league, one stat is as valuable as the other and it will leave you with a hole in your roster.

    Vote -1 Vote +1

  13. rubesandbabes says:

    Something that hasn’t been mentioned is that very few guys really rack up the holds each year – the curve flattens out very fast.

    Suggested: American League 8th inning guy on a winning team.

    Vote -1 Vote +1

  14. Matt says:

    awesome!

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>