Welcome back for Part 6 of The Eighteenth Out. (Part 1, Part 2, Part 3, Part 4, Part 5) Every article has brought us a step closer to redefining a new, effective way to go about evaluating and drafting point pitchers. Thankless is the job of the Point League fantasy analyst, yet we trudge on. There are times in life we break away from the pack and wander out alone into the unknown, determined to achieve our destiny. Imagine something that epic, but relating to baseball statistics (and a full orchestra playing the GoT theme song in the background). Why so serious you ask? You’ll see (I hate being wrong).
Not a single turn in our venture to dissect, and further understand the Quality Start’s place in the landscape of imaginary baseball has come easy. We have battled detest from an industry that doesn’t even care to keep their QS statistics correctly (For the record, those bogus QS stats I drew from are still up). Then even the most rudimentary of sorted searches proved futile, and we had to create our own custom stats. It hasn’t all been pretty. It hasn’t all been without flaw. Failing 70% of the time has gotten players into the Hall of Fame; I’m not quite sure the same leniency is extended to analysts. This is my first experience writing. To me, it’s an immersive exploration; And we’re figuring this stuff out together as a collective. I make mistakes like anyone else, but unlike everyone else I own them and correct them. I mined poor data, and now have had to revisit some of the correlative data from Part 4. After doing some testing I assure you the integrity of the assertion remains intact, as is evidenced in the scatter chart, and furthered after transposing the QSW data. However the values will all shift a bit, albeit in the same direction. I didn’t want to leave that unaddressed, but we deal with the BS when it comes our way. And we do it happily. Why? Well, if you have ever lifted a home league trophy and used your friends’ hard earned money to go to the Bahamas, you know exactly why.
Back to business. At this stage we’re getting far enough into our work I feel a quick recap will prove useful before diving right back into the numbers. Thus far we’ve poked and prodded at the strength of achievement associated with a Quality Start to discover quite the dichotomy; The presence of a weak link to rotisserie results (mostly based on the potential for a QS w/poor per inning efficacy), yet simultaneously a strong correlation to point finish. We’ve identified the disproportionate reward for the eighteenth out as a cause, but not the only factor for this. Our desire to exploit the pertinence of this relationship drove us to create and cultivate an entirely new and critical stat for our purposes; The Quality Start + Win (QSW). We’ve also connected 2018 QSW to point finish by position rank, and, most recently identified players with unsustainable Quality Start + Win Conversion Percentages (QSWC%) as candidates for win regression in 2019.
Regression candidates are always a bit of a bummer. having some random guy take shots at your favorite deep sleeper. Let’s brighten things up a little and look at the inverse of that same Quality Start + Win Conversion Percentage chart (Custom stats found only here at Razzball by @MLBMovingAvg; Tell your friends). This time we’ll look at the Starting Pitchers who in theory stand to see an uptick in 2019 QSC%. SPOILER ALERT: Since we already tied the QSC% leaderboard to top bullpens or teams w/a .580+ win pct, we will be applying those same forces here. Personally I will be looking for regression to the mean for any extreme outliers in either direction, weighted for the quality of the team of course.
I wanted to make up for my previous bloopers with a bit of generosity. Here is the entire list of Starting Pitchers in 2018 with a minimum of 5 QS that won less than 50% of their Quality Starts. Keep in mind the league average is right around a 55% conversion rate.
| NAME | QS | QSW | QSWC% |
|---|---|---|---|
| Bailey, Homer LAD | 6 | 0 | 0.00% |
| Fister, Doug TEX | 5 | 0 | 0.00% |
| Fulmer, Michael DET | 10 | 1 | 10.00% |
| Hammel, Jason KC | 7 | 1 | 14.29% |
| Matz, Steven NYM | 12 | 2 | 16.67% |
| Giolito, Lucas CWS | 15 | 3 | 20.00% |
| Guerra, Júnior MIL | 10 | 2 | 20.00% |
| Cashner, Andrew BAL | 13 | 3 | 23.08% |
| Archer, Chris PIT | 12 | 3 | 25.00% |
| Holland, Derek SF | 11 | 3 | 27.27% |
| Borucki, Ryan TOR | 11 | 3 | 27.27% |
| Rodriguez, Dereck SF | 14 | 4 | 28.57% |
| Bumgarner, Madison SF | 14 | 4 | 28.57% |
| Pena, Felix LAA | 7 | 2 | 28.57% |
| Liriano, Francisco DET | 10 | 3 | 30.00% |
| López, Reynaldo CWS | 19 | 6 | 31.58% |
| Shields, James CWS | 19 | 6 | 31.58% |
| deGrom, Jacob NYM | 28 | 9 | 32.14% |
| Cobb, Alex BAL | 15 | 5 | 33.33% |
| Kennedy, Ian KC | 9 | 3 | 33.33% |
| Sánchez, Aarón TOR | 9 | 3 | 33.33% |
| Lauer, Eric SD | 6 | 2 | 33.33% |
| Hamels, Cole CHC | 16 | 6 | 37.50% |
| Anderson, Tyler COL | 16 | 6 | 37.50% |
| Weaver, Luke ARI | 8 | 3 | 37.50% |
| Hess, David BAL | 8 | 3 | 37.50% |
| Buehler, Walker LAD | 13 | 5 | 38.46% |
| LeBlanc, Wade SEA | 10 | 4 | 40.00% |
| Chen, Wei-Yin MIA | 10 | 4 | 40.00% |
| Germán, Domingo NYY | 5 | 2 | 40.00% |
| Teherán, Julio ATL | 17 | 7 | 41.18% |
| Sánchez, Anibal WSH | 12 | 5 | 41.67% |
| Ross, Tyson DET | 12 | 5 | 41.67% |
| Montgomery, Mike CHC | 7 | 3 | 42.86% |
| Bettis, Chad COL | 7 | 3 | 42.86% |
| Gausman, Kevin ATL | 16 | 7 | 43.75% |
| Wheeler, Zack NYM | 18 | 8 | 44.44% |
| Leake, Mike SEA | 18 | 8 | 44.44% |
| Pivetta, Nick PHI | 9 | 4 | 44.44% |
| Straily, Dan MIA | 9 | 4 | 44.44% |
| Kershaw, Clayton LAD | 20 | 9 | 45.00% |
| Suarez, Andrew SF | 11 | 5 | 45.45% |
| Estrada, Marco TOR | 11 | 5 | 45.45% |
| Wood, Alex CIN | 13 | 6 | 46.15% |
| Bundy, Dylan BAL | 15 | 7 | 46.67% |
| Heaney, Andrew LAA | 17 | 8 | 47.06% |
| Roark, Tanner CIN | 17 | 8 | 47.06% |
There were 47 Starting Pitchers in 2018 with a minimum of 5 QS that won less than 50% of their Quality Starts. Of those 47, 17 pitchers (36% of total) were on teams who ranked in bottom 10 in bullpen ERA during the 2018 season. 21 of the remaining 30 (44.6% of total) played for teams without a winning record, including .500 teams. Therefore 39, or 80.8% of these 47 starters could link their poor QSWC% to those aforementioned criteria. The effect that team win % and bullpen strength has on QSWC% (Though initially intuitive) rears its ugly head once more in a manner we are beginning to quantify. If time allows, at some point I’d like to circle back and look for notable links to 3rd, 4th & 5th starters on good teams in later rounds compared to 1st & 2nd starters on bad teams that get drafted early. Let’s keep that plate spinning and get back on task. Theoretically, a repeat performance from any of these outlying pitchers should yield a better win conversion percentage in 2019; In particular those 9 who played for teams with a top 10 bullpen or a +.500 squad. I list them below, with the adjusted wins accounting for a correction to the conversion mean. Keep in mind here that we saw many of the outlying players to the positive side that played for better teams not only get back to the average, but greatly exceed it. To cover this very possible outcome in our expected range I’ll also include the adjusted wins if they were to progress to a 75% QSWC%.
- Guerra, Junior +4 to +6 adjusted Wins
- Hamels, Cole +3 to +6 adjusted Wins
- Buehler, Walker +2 to +5 adjusted Wins
- LeBlanc, Wade +2 to +4 adjusted Wins
- German, Domingo +1 to +2 adjusted Wins
- Teheran, Julio +3 to +6 adjusted Wins
- Montgomery, Mike +1 to +2 adjusted Wins
- Gausman, Kevin +2 to +5 adjusted Wins
- Kershaw, Clayton +2 to +6 adjusted Wins
These theoretical wins lost to the headache that is relief pitching are absolutely worth considering in our valuations going forward regardless of format. So if any ardent Rotisserie players happen to still be with us, this Bud’s for you. I have yet to see this angle taken towards win adjustment anywhere else, so any other advancements here would continue our trend of pioneering work in the fantasy space. We’ve used our own custom stats to develop a new lens with which to project, but can we take it further? If you aren’t already strumming an air guitar and screaming ”HELL YEAH BROTHER!”, there’s a problem.
As promised in coming segments of the series, we will really put the screws to our discovery’s usefulness. We’ll take these newly cultivated custom stats and split them home and away. My hopes here are to identify pitchers with very heavily varying splits that make 2018 total season stats look unappealing, and create potential player pairs to alternate throughout the season. Before we tackle that dummy, I want to leave that plate spinning (another Part 4 reference) for a little while and pivot towards something different. I couldn’t help thinking that as we further prove a Quality Start on its own to be a bit wonky, that clumping all different varieties of QS together was distorting the clarity of our results. So, for your enjoyment we’ll get out our magnifying glasses once more and see how that QSWC% changes with differing QS results (all scored the same, mind you).
And so concludes another leg of this fruitful journey with your favorite friendly, neighborhood fantasy analyst John L. @MLBMovingAvg. I would never had guessed at the outset that I would stumble on so many fascinating and pertinent relationships otherwise overlooked when doing point pitcher evaluations. Thanks for reading, following, responding, correcting (hopefully for the last time) and supporting me at Razzball. C’mon fellow explorers! More comments! More feedback! We wont stop until I’m invited on the podcast with the legend himself!
