The value of not being broke

The minor hubbub yesterday during the draft when Atlanta was forced to make a worse selection than they otherwise would have because of a lack of funds to pay a signing bonus put me to thinking. What is the overall difference in potential value between the pitcher that they did draft, Orlando Andrade, versus the pitcher they could have drafted who was subsequently selected by St Louis, James Mitchell? It also gave me a 15 minute break from end of quarter spreadsheets, while allowing me to screw around in a spreadsheet so it didn't look like I was screwing around.

A couple of big caveats ... I was bored enough to not bother looking at standard evaluations and instead wandered along my own random thought process. I ended up deciding to look at how effective a starting pitcher was at suppressing runs scored based on the number of baserunners allowed, i.e. how good they are at assisting their defense in preventing runs from scoring once they reach base? Second, I was too lazy to dig up very detailed stats, so I used just a basic set; IP, hits, runs (not earned just total runs), homeruns allowed, walks, and strikeouts. Third, I do believe that pitchers can influence the outcome of balls put in play. See this article from Diamond Mind Baseball for a detailed analysis on why this is true, even if it is usually a very minor difference between one average pitcher and the next. Without going into too much detail on the contents of the article, I will point out that the difference between a good pitcher and an average pitcher may be small, but an average pitcher is actually going to be a pretty good pitcher. Average pitchers stick around, have prosperous careers, and are valuable. What we really want to look at is how much of a difference can a pitcher provide over a replacement level player, or that VORP stat that v6.5 doesn't calculate. This was slightly difficult as most NOBL teams didn't really have any replacement level starting pitchers on their 2017 staffs that saw significant playing time in 2016 from which I could calculate the baseline stats, but luckily for you I persevered and just aggregated up a couple of duds who shall remain nameless to save their GMs the embarrassment.

On to the analysis, after which you can all feel free to point out the severe flaws in my logic. Before I could look at the actual value for two pitchers who probably won't even make a major league appearance for another 3 or 4 seasons, I had to find similar contemporaries with enough data to crank through. Enter Alvin Aybar and Dustin Harrison. Neither will be an exact match to Andrade and Mitchell, respectively, but they are pretty damn close; Andrade will probably be not quite as good as Aybar, and Mitchell will potentially have much more movement on his pitches than Harrison which could have a large impact on improving his chances. As an after thought, I added World Series MVP Michael Drexler to the calculations as another potential match for Mitchell. Their stuff and movement are close, but Drexler does have control issues which hurt his overall stats. Mitchell is probably going to be a mix of Drexler and Harrison; Drexler's nastyness with stuff that offers plenty of movement, but the slightly better control of Harrison. Of course, this entire premise is based on Andrade and Mitchell both actually developing to their talents, but why not ... it could happen. To round out the eval, I then aggregated a couple of our duds, and added in Yoo Ki Son as well just to see what an extremely dominant pitcher's stats would look like.

To start with, I did a quick and dirty pro-rating of every one's stats to equalize all innings pitched while maintaining the respective starters personal ratios for the remainder of the data. I used Aybar's 1622 innings pitched as the baseline, largely because until I added Drexler at the last minute, he had the most innings pitched. This extrapolation does potentially negatively effect Harrison, who as a younger pitcher will be more adversely effected by his youthful wildness whereas his total walks when he does reach 8 seasons of his own work may be lower than the projection, but probably not by much. Aybar is also a safe bet as the baseline as he has not seen a significant drop off in performance in his later years, so he has a good solid 8 seasons of reliably consistent ball when it averages out.

Equalized on innings pitched, the stats look pretty much as would be expected. Aybar surrenders fewer runs and homeruns than the Replacement Pitchers (RPs) but more than Harrison. He did actually give up more hits than anyone else, more than one per inning, but made up for that with a very good walk/inning ratio that was better than Harrison by a long ways. The RP pitchers are appreciably worse in every category, although they did strike out more hitters than Aybar by a negligible amount, and Ki Son was significant better in every area than the others.

OK, diving in ... what I really wanted to look at was on average, how many outs does their defense record compared to the number of runners that are actually scoring. I removed homeruns and strikeouts from the calculations because they don't impact what I'm looking for; the pitchers interaction with their defense. A run is guaranteed to score on a homerun, and other than the extremely rare circumstances of an outfielder turning a homerun into an out or the inside the park homerun, there is no opportunity for defensive interaction on the longball. I am looking primarily at hits plus walks minus homeruns, and how that compares to runs allowed minus homeruns over the 4866 total outs minus the strikeouts presented in our rationalized 1622 innings. It would seem that these calculations would give a reasonable representation of 1) how good the pitcher is at avoiding successive baserunners (which WHIP also covers), 2) how good the pitcher is at avoiding extra base hits (other than HR) which are proportionally more likely to lead to runs, and 3) how good they are at putting balls in play within range of their defenders. All of this assumes that these ratios are based on skill as opposed to luck or excessively great defenses, which will always play into it as well but the better pitchers should have better ratios.

The RPers are as bad as could be expected at recording defensive outs per run. They only convert 1.7 defensive outs for every baserunner they allow, and are only retiring 4.85 hitters for every runner that crosses the plate. They are, if the data set means anything like what I'd hope that it means, really bad at helping their defense out. They put too many men on base via the free pass and balls in play manage to elude defenders. Not shocking in any way, but it allows us to set the baseline.

Aybar is appreciably better at recording outs in lieu of allowing the batter to reach, with defensive outs at 2.17/baserunner and 6.02/run. Every third batter still results in someone standing on a bag scratching themselves inappropriately, but they score much less frequently than a RP. A full 27% improvement in baserunners, and a 24% improvement in runs allowed per defensive out.

Harrison is actually worse than Aybar at preventing baserunners per defensive out, entirely because of the free passes. It's hard for your defensive to work effectively when they are watching batters stand there with the bat on their shoulder. Harrison generated only 1.83 defensive outs per baserunner, which is still better than the wilder Drexler who could only muster a 1.73 ratio. Harrison is only 8% better than the RP pitchers in this aspect, but is a massive 45% better than the RPs at preventing those runners from scoring, generating 7.07 defensive outs per run. Harrison is almost twice again as effective as Aybar at allowing his defense to get him out of his own jams. He might put more people on base, but when batters do put the ball in play, his stuff is keeping them from driving runners home.

So what does this potentially mean for Atlanta, when we transpose these numbers on to the future Andrade and Mitchell? Andrade is not going to be a bad pitcher by any means. Significantly better than a replacement level pitcher no matter how you look at it, and much more congenial with his defense. However, the same nasty stuff that makes Mitchell so much more effective at two things I didn't loop in, generating strikeouts and preventing homeruns, also will allow him to make life much easier on his defense. In a game of trading outs for runs, Mitchell will have an extra defensive out in hand for every run allowed, or 2/3 a run per 27 outs. If every ball were put into play with a defensive opportunity, Mitchell would be need 15% less offensive support per game, allowing only 3.8 runs per 27 outs to Andrade's 4.49. 10 runs a season in defensive outs alone, down the tubes because they couldn't afford a signing bonus.