Streaming for W's and K's...Are we doing it wrong?

   Back to Forum Search Topics Edit User Settings Scroll to Bottom


 You can add an avatar image from your User Settings page.

avatartjabshire (7 posts)  3/20/2014 6:14 AM

Streaming back of the rotation starters has always been the standard method of trying to steal some wins and strikeouts. Even in leagues with a limit in the number of innings you can play, the best way to compete in these categories is by coming as close to that cap as possible. The reality of it is, everyday there are wins going to unrostered guys. My hypothesis (and I'd love to get some help from somebody more statistically inclined) is that filling every possible P spot with a middle reliever is the more efficient method of picking up wins and K's, especially if you want to keep strong in ERA and WHIP.

The first core concept of this is that since there is a limit on innings pitched, and we intend to reach that limit, all counting stats should be judged on a per inning pitched basis. The traditional approach has been to trade approx. 6 innings from 1 guy for a shot at 1 win. Instead those 6 innings are a scarce resource that can be spent on 6 guys pitching one inning each. One might think that starters pick up the wins too frequently for you to still gain this way so I'll provide a case study: Player A is projected to get 17 wins in 232.2 innings, while Player B will get 4 wins in 40.8 innings. That works out to roughly .073 wins/IP for player A and .098 wins/IP for player B. Working with a hypothetical limit of 1000 IP for the season, that would come out to 73 wins on the season for a roster in which every IP was thrown by Player A or a pitcher of similar caliber, while a team of Player B's would win 98 games.

Seems easy right? Player A is Clayton Kershaw, Player B is Cory Gearrin. You could have every allowable inning pitched thrown by the best starting pitcher of this generation and still lose handedly to a team of guys who you could get off of the waiver wire in all but the deepest of deep leagues.

While Gearrin > Kershaw in wins/ inning, it has to be noted that Kershaw still has the advantage in ERA, WHIP, and K/9. The point of writing this isn't to say the last pitcher on the page is better than the guy listed first. The thing to take away from this is that there are a lot of better relievers than Cory Gearrin available on waivers and that everybody else is going to be starting pitchers with much worse numbers than Clayton Kershaw.

Only 11 of the top 40 pitchers in era are starting pitchers. For WHIP it's 16. Many of the relievers who crack those lists are unowned in most leagues. When we move into the counting stats -- keeping in mind our new approach to counting them on a per inning basis-- the numbers are even more telling. Of the top 40 pitchers in W/inning only 6 are starters; meanwhile,  K/9 is where are the flame throwers in the pen compose a whopping 38 of the top 40 spots.

That's four categories so far in which top guys are all available in abundance, meanwhile names further down the list populate your opponents' rosters just because they start. That's before we've even begun to consider that 5th category in a typical 5x5 roto league. It's so obvious, and easy to overlook, but let's not forget that 40 out of the top 40 in Saves are relief pitchers. Streaming a starting pitcher is a conscious choice that you will spend a few of your precious limited innings with absolutely 0% chance of possibly getting a save out of them. Obviously, few leagues have closers sitting around as free agents, but competing for counting stats over the course of a season is a war of attrition and you'll be the one picking up the random saves throughout the year to help bolster that final total. Who knows? You could even find yourself holding a guy who's about to step into the closers role giving you a huge advantage out of nowhere. You've given yourself considerably better odds than everybody else of finding that guy.

Now that the superiority of the RP has been demonstrated over the course of an equal number of innings, we're faced with the challenge of playing as many RP innings as possible while most importantly making sure that the maximum inning pitched limit is reached. While the top 40 names on projections for each category were dominated by relievers, the starters who made the lists were typically the same names over and over. There are probably less than 10 starting pitchers whose numbers are consistently better than RP's available on waivers. You want as many of your innings pitched as possible to be pitched by either those elite pitchers or closers. Every other inning between whatever your stars and closers play and the cap should in theory be played by a RP.

The second core concept of this plan is that you need to try and get as many innings out of your relievers as possible. If a guy has a day off, cut him for somebody who is playing. If a guy threw yesterday, cut him for somebody rested. Bench your starters and closers when they're not playing and play a reliever, using dual eligibility players to fulfill any league roster requirements.  You need every possible starting spot filled with somebody who could throw each day. The waiver wire is full of guys that offer more per inning production than 90% of the starters on your opponents' rosters. Remember, if you match them inning for inning, you'll blow them away in 5 categories. Getting to the max innings will take work.

I think the math speaks for itself that this is the proper way to manage a team for ideal success in leagues with unlimited moves and limited innings pitched. The only question that remains is how many innings would ultimately be attainable. I can guarantee that guys with good per inning numbers who have a chance to pitch on a daily basis will be available on waivers, but there is no certainty that the player you've picked will actually get in the game. Is there anybody out there with any experience with this kind of thing? I'm especially interested if anybody had any clue how many innings per starting spot you could expect to accrue. Are you able to pick up on bullpen usage patterns enough to average over half an inning played from each starting spot on a daily basis? There are 188 days with games played this year. If you could get half an inning each of those days, that's 94 innings for each starting spot. I feel like you could possibly get even more than that by learning how managers use their pens. Is there a standard for how many innings leag


0% Agree (0 votes)
avatarliam4 (283 posts)  3/20/2014 10:54 AM

Also of note -- I believe the z-scores only apply based on the size of your league.  i.e., if you have 10 teams with 15 rostered hitters each, it'll take those top 150 and calculate averages and standard deviations from there.

Ken, on this thread, would it be possible to add a W/9 league category?


0% Agree (0 votes)
avatarsilentjim (54 posts)  3/20/2014 11:47 AM

I don't play in any leagues with an IP cap, just a minimum, so I agree with completely.

I've mentioned in a few other threads how the wins catergory is blown wildly out of proportion based on middle relievers and closers vulturing (or earching) some. It's once of the reasons I like QS.

Anyways it's hard to look at past leagues actual stats since I can just see Yahoo final roster stats, but in 2011 the last season my one competitive league used wins, the leader of wins had 89 total. His final roster has 107 total (trades, not starting guys, pickups I assume). Of the 107, 22 of those wins came from relief pitchers. 22 of 89 is 25% or 22 of 107 is 21%. Or if you try to extrapolate back you'd get 18% of wins came from RP of the 89 (small sample of course). That was also out of 312 IP and nice peripheral stats (303Ks, good whip and ERA). This was 5 MR and closers total (Javy Guerra, Mark Melancon, Bobby Parnell, JJ Putz, and Drew Storen)

Years ago people always talked about how three good middle relievers good equal an ace especially with the inconsistency in wins as a cat.


0% Agree (0 votes)
avatartjabshire (7 posts)  3/20/2014 1:52 PM

I was writing that pretty late, and just realized all of my numbers were NL only. Mixed leagues would be even easier to find a couple stud SP and a doubling of the available reliever pool. The same concepts would still hold true.


0% Agree (0 votes)
avatarRapierman (72 posts)  3/20/2014 4:41 PM

Two more things to consider:  1.  Trends, and 2.  Situations.  Is one pitcher better against this batter in this park at this time than another pitcher against another batter in another park?  Is there someone on a hot streak that's useful enough to take advantage?  These things must be considered when you're streaming.


0% Agree (0 votes)
avatarmbuser (772 posts)  3/21/2014 12:44 AM

Obviously we're dealing with roto here, and I've been a proponent of heavily utilization of MR in roto leagues for a long time now. To what you said, Jim, combined Tyler Clippard, Drew Smyly, and Tanner Scheppers last season and you get the following line:

222.7 IP, 18 W, 3 SV, 2.22 ERA, 0.99 WHIP, 211 K


0% Agree (0 votes)

    Scroll to Top