Fantasy Baseball Streaming – Valuing Starting Pitchers

This post leverages Rudy and Grey’s Razzball Commenter League experience as well as some modeling (the dorky kind) to quantify the effectiveness of streaming and how it should inform one’s draft strategy in shallow mixed leagues (10-12 teams).  The streaming decisions made by Grey and I were HIGHLY influenced by our free, daily updated tools for streaming starting pitchers (Stream-o-nator) and hitters (Hitter-tron).

If you are reading this article (or this site for that matter), I assume you are familiar with streaming starting pitchers.  This is an essential strategy in all daily league formats whether one plays standard Roto or H2H.  While my beloved Stream-o-nator aims to make sure all our readers make the most informed decisions on which pitchers to stream over the next 7 days (and possibly more in 2014), I have yet to read any analysis that quantifies the value of the average streaming pitcher to inform draft strategy.  So this post is going to focus on quantifying the value of a streaming SP and, once I’ve completed quantifying the value of streaming relievers and hitters, I will figure out how this impacts draft strategy.

For this analysis, I am focusing on 12-team 5×5 mixed leagues with moderate GS/IP caps (~180 GS/1,350 or less IP).  In this format, most teams are going to hit their GS/IP cap so the value of a streamed GS is how it compares to a rostered pitcher’s GS.

I will try to directionally speak to the impact in other formats including:  1) large GS/IP cap leagues , 2) leagues with weekly roster changes, and 3) H2H leagues.  Regarding #1 and #3, the value of a streamed GS can be higher as it provides surplus W/K (versus no GS) and their value is higher than the ERA/WHIP downside.

One extra caveat – while ‘streaming’ implies the owner is constantly churning through pitchers, I define it differently.  I see it as pitchers that are picked up off free agency/waivers during the season.  So if I choose to hold onto that pitcher for a stretch, that is still streaming to me.  The important point is that these are pitchers that were neither drafted by your team nor were acquired via trade.

Valuing The ‘Composite Streaming Pitcher’ For 12-team Moderate GS cap leagues

Based on 2 years of using the Stream-o-nator, it seems like the projected value of the best available SPs in 12-team leagues is usually in the $6-$10 range.  So my expectation going in is that this is the ‘high’ range for the value of a composite streamer (i.e., combining one’s streams so that it has the same GS as a standard pitcher).  I would also imagine that the more one streamed, the lower this composite value would be since they could not afford to be as discriminating.

To test this, I went into the active stats of 3 RCLers who did significant streaming:  myself, Grey, and Retro Vertigo who finished tied for 2nd in the overall RCL standings and competed in the most competitive league (ECFBL).  I includes the stats of all starting pitchers that 1) were not drafted and 2) were not traded for.

As you can see, all three of us were in a similar range (426-636 IP) with Grey on the high end and me on the low end.  This was mostly out of necessity as I had a more successful SP draft than Grey (with 80+ GS and about 600 IP going to Cliff Lee, F-Her, and Derek Holland).  My average innings per non-drafted starter (IP/SP) was the highest among the three because a large chunk of my streaming IP (a little less than 50%) was invested in Justin Masterson (120 IP) and Dillon Gee (83 IP).  Grey and Retro Vertigo didn’t go as bestie on streamers as me with their highest IPs being Grey’s Gerrit Cole (37 IP) and RV’s Chris Tillman (36 IP).

To determine the $ value and to make easier comparisons between the three lines, I pro-rated the stats down to the average starter’s GS.  Since ESPN does not break out GS in the ‘Active Stats’, I went with the assumption that a streaming starter delivers 6/7th the IP as a typically rostered SP (think 6 IP/GS instead of 7 IP/GS).  I have the average SP at 197.3 IP so the pro-rated IP total is 169.1.

Retro Vertigo did a slightly better job than Grey and I at streaming last year – although the difference is minor enough (about $3 when multiplied out to the total streamed IP) that Grey and I’s better performance (or preference) for Wins/K over ERA/WHIP might have been better for us in term of standings points.  (Note:  My Masterson/Gee combo were worth $14…meaning my other streams were worth around $3.  Ugh.)

So it does appear that $6-$10 is a solid range for ‘best case’ streaming composite pitcher value in a 12-team league (we are talking about the guy who runs the site and RV finished in the top 1%).  My $3 streaming value (sans Masterson/Gee) shines some light on the less glamorous range of potential outcomes.

In a GS cap league, however, there is a bonus value to streaming SPs – you can plug in a middle reliever into the roster slot while between streams.  If you figure that you only need 30 GS out of your designated ‘Streaming SP’ roster slot, then you have 130 games or so to slot in middle relievers.  Assuming a competitive league (where you need to grab the SP a day or two before gametime) and cases where one holds onto a pitcher between starts because of a favorable matchup, let’s reduce that 130 games to about 60 games.

Below are the pro-rated stats for 20 innings of non-drafted middle relievers – using Retro Vertigo’s stats since Grey’s streamed relievers were heavy on SAGNOF and I had so much drafted RP value (including K-Rob and 1/2 a year of Jansen as MR) that I had little need/room for it.

How is this worth almost $4?  Well, adding this MR’s stats to an average team has no impact on ERA/WHIP but represents 0.2 Point Shares for Wins, 0.2 Point Shares for Saves and 0.6 Point Shares for K’s (i.e., adding 20.7 K’s, on average, would make a team with a 10 in K’s go to a 10.6).  An additional Point Share is worth about $3.9 and this composite pitcher is worth about that.

(I ran the same analysis using Grey’s SAGNOF-heavy MRs and it would’ve netted $4.6 – trading off some ERA/WHIP for 3 more Saves)

So if one is a dedicated and good enough streamer, the value of that roster spot is around $12 ($8+$3.7).  Using my $3 as the ‘floor’, I’ll put the range at $6-$12 for this roster spot with $9 as the ‘average’.

Valuing The ‘Composite Streaming Pitcher’ For 12-team Moderate IP cap league

The difference between a GS cap vs. IP cap league is that the rostered SP’s advantage in IP per GS no longer matters.  This increases streamer pitcher value as it removes the IP/GS advantage that the rostered SP has (which reduces streamer K potential by 14%) and increases the number of GS you can cram into the IP cap (which increases the Win potential for streamers to 17% vs a rostered SP).

If the above gave you a headache, hopefully the table below will make more sense.

Here are the same stats from the earlier section but pro-rated to the average IP total I used for a rostered starter:

So the value of a streaming SP in a moderate IP cap league is about the same as in a GS cap league as the value of those extra streamer SP innings (in this case) and the streamed MR innings (in the GS cap case) end up balancing out.

Valuing The ‘Composite Streaming Pitcher’ For 12-team Large GS/IP Cap Leagues

In a large GS/IP cap league (including ones where there are no cap), it is possible that not every team maxes out their pitching.  A streamed start in that context means that all Wins/K’s are additive and the only possible downside is on ERA/WHIP.  Given that ESPN/Yahoo/CBS league defaults have GS/IP caps, my guess is that large/no GS/IP cap leagues are disproportionately deeper leagues and have weekly vs daily roster changes (this is the case with every 15-team MLB and AL/NL-only league).

So the value of a streaming pitcher when replacing a spot normally dedicated to a drafted SP is the same as in moderate GS/IP cap leagues.

To estimate the impact of additive GS in this type of format, I added my pro-rated SP values to the average SP line for 12-team leagues – moving the IP from 1,361 to 1,530.  If somehow you were able to stream at an $8 level AND pitch 170 more IP than the average team in your league, the value is around $25 with gains of 4 standings points in both Wins/Ks while only giving up a point each in ERA/WHIP.  You can also look at as you are gaining $1 for every 6.7 IP.  Even if you can stream at a $4 level, you would be looking at $1 about every 15 IP.

Note:  It’s also possible that in a large GS/IP cap that there are more SPs rostered to max Wins/Ks which would might make it harder to find quality streaming candidates.  Net-net, maxing your GS/IP is almost always a smart strategy as long as you are adding pitching starts that are above $0 and maybe even a few dollars in the negative (as it is guaranteed upside on Wins/K’s with, on average, lesser downside on ERA/WHIP)

Valuing The ‘Composite Streaming Pitcher’ For 12-team H2H Leagues And/Or Weekly Roster Change 

In both these roster formats, the common streaming practice is to target the ‘2 start’ pitchers to maximize counting stats.  Everything equal, maxing GS/IP is a positive.  Given everyone is on the search for 2-start pitchers, you are typically left either 1) taking a crappier ‘2 start’ pitcher than the average available ‘1 start’ pitcher or 2) rostering a lot of SPs.

So the fundamental valuing question in this type of format is really “How much worse does a 2-start pitcher have to be than a 1-start pitcher that their ERA/WHIP impact hurts my team more than their W/K contribution?”

To model this, I looked at Stream-o-Nator data for starts worth exactly $1-$10 and averaged them together for a composite.  It looked like this:

I then added these totals as well as 2x these totals (to reflect 2-start pitchers) to the average team’s yearly totals to estimate their impact.  The regressed formula I got from looking at 2-start pitchers with average start values between $2-$6 is the following:

-1.635+2.125*(Average of 2-start $ Value)

Here is how that translates:

The reason that a 2-start pitcher is worth less than its sum (e.g., two $2 starts not worth $4) is because the impact on Wins/K’s is slightly below 2x (since a $4 pitcher has slightly better Win/K than a $2) and the two-start pitcher drags down ERA/WHIP more than 2x (because the $2 pitcher has worse ERA/WHIP than a $4).  This should hit the 2x mark when the `2-start pitcher` average $ value equals the value of the average GS.