On Tuesday, I provided the general approach to my personal positional rankings (which you can find in Tiered form on Rotobanter), positional scarcity adjustments using the "replacement level" at each position, and associating auction values (to z-scores) using Zach Sanders' Fantasy Value Above Replacement series on FanGraphs.
Yesterday, I presented the hitter projections with Freddie Freeman as my example. Today, I'll focus on pitchers. This is all to support Ray's schedule for the 2014 Fake Teams Consensus Position Rankings, which starts this Monday...Gladiators Ready? Contestants Ready?
If you prefer to look at our auction values in terms of a batting-average like value (.300-like scale), check out Daniel Kelly's Fantasy Average Positional Breakdown.
Onto the approach:
IP & TBF (Total Batters Faced): I simply look at the career average/trending IP per GS and TBF per IP, naturally multiply them by GS and IP respectively, and we have our Innings Pitched and Total Batters Faced output.
K/BB: For command-related data (K and BB), we get our K and BB total by multiplying their K & BB% to TBF.
Balls In Play Mix, HR/FB & WHIP: Fun starts here. Similar to K & BB% which is projected somewhat manually (combination of a weighted 3-year average, trends and somewhat of a manual pitch f/x incorporation this year), we land on their GB and FB data (IFFB and HR/FB as well). Groundballs and Flyballs correlate relatively well year to year. Their LD% comes from 1.00-GB%-FB%. Now that we have these percentages, we can multiply each one by our quantity of Balls In Play (TBF-K&BB).
Example: Clayton Kershaw
32 Games Started * 7.15 IP/GS = 228.8 IP
3.89 TBF/IP * 228.8 IP = 890 TBF
25.5% K-rate & 5.9% BB-rate = 227 K's and 53 BB's (4 HBP's)...combined with TBF and we're at about 607 Balls In Play.
46.3% GB-rate; 33% FB-rate (11% IFFB-rate) & 20.8% LD-rate...combined with BIP and we're at about 280 GB's, 200 FB's and 125 Line Drives. Through hit expectancy (% of each Balls In Play Type that become hits), we land on 182ish hits, however we now account for Hit-related Park Factors, which by the way is relatively glorious in LAD and 3/5 stadiums in that division. After the Park Factors incorporation we're below 180 hits.
If you're in the "Holy Trinity" of pitching, we now have his K/BB (4.32) and GB/FB ratios (1.4) as well as his projected WHIP of 1.01
...2/3 Fantasy 5x5 Stats down: K & WHIP...onto Wins and ERA.
Real quick on Wins:
One of the better ways to project wins is to have all your projections complete and then understand that pitcher's runs per 9, but I need to project wins throughout the offseason for posts prior to completion of every pitcher and every teams' expected runs total. I therefore use a simpler approach accounting for each pitcher's career wins per ER and IP. Kershaw's ratio is 5.25 which gets us to 15 Wins this year. Scherzer on the other hand has a 4.89 ratio but 16 wins contingent on his historical (at least Detroit-related expectancy). For reference, the only other ratio I have above 5.00 (5.08) is Adam Wainwright which associates to a 17 Wins projection.
Hopefully you're still with me here. This might be the segment that would provoke the most debate if you are still with me. There are different expected ERA's available weighing certain skill-sets differently. I'll let you search for xFIP and SIERA on your own, but my new favorite (so long as you project all the incorporated outputs) is Stephen Staude's BERA. In short, here is the formula with the shortened version as well:
BERA = (32.6*HR + 11.4*BB – 7.9*K)/TBF + (5*LD + OFFB – 3.4*IFFB)/(FB+LD+GB) + 2.2*ZC% + 0.22*SIP% + 0.51
11.4*BB% – 7.9*K% + 5*LD% + FB%*(1 – 4.4*IFFB%) + 32.6*(HR/TBF) + 2.2*ZC% + 0.22*SIP% + 0.51
This is what I use as my ERA base, but I then incorporate three additional factors - one is obviously Park Factors. Another is one that Stephen Staude agreed he wants to add (SwStr%). The last is LOB%. LOB% is pretty fluky year to year, but it has a ton of effect naturally on ERA so it's important to factor in.
Here are Kershaw's expected ERA's after park factors adjustments:
Expected ERA: 2.13
and finally, BERA (as my base) with my inclusion of SwStr% and LOB%: 2.59
I have Kershaw as follows: 15 Wins - 2.59 ERA - 1.01 WHIP - 227 K. His non-adjusted (for position scarcity) Fantasy Value (z-sum) = 11.826, which puts him third overall:
Mike Trout - 13.23
Miguel Cabrera - 12.43
Clayton Kershaw - 11.83
Paul Goldschmidt - 9.9
Andrew McCutchen - 9.3
*These are not $ values, but their Fantasy Values (z-sums from their 5x5 z-scores)
The big upgrade to next years' Pitcher projections will be the further incorporation of Pitch F/X - understanding the pitch efficacy on these sabermetric outcomes (BIP mix and SwStr%). To see what I mean, you can check out my Pitch Score post I put up on Beyond the Boxscore.
This post focused primarily on the 5x5 output...how I came up with their W, ERA, WHIP & K (I don't project Saves for relievers - it's most about opportunity then skill), but if you want to understand the weighted 3 year averages a bit more of stats/peripherals that correlate well year to year (regression to the mean/trends, etc.), you can comment or send me a request.