RADIO
When we talk about a 'clutch' baseball player, there are various ways we can describe them. A pitcher who can get the big out or a hitter who can come through and get the timely hit, on at least a semi-regular basis.
Generally speaking, a 'clutch' player is someone who will be his best in a high-leverage situation. But what are those high-leverage situations?
If we look at pitchers, what would be the scenario that best exemplifies their 'clutch' characteristics?
Featured Promo: Get any MLB Premium Pass FREE ($119 value) AND up to a $200 deposit match when you sign up with Rival Fantasy. Rival offers 6 and 12-team Best Ball, weekly and daily contests, for as little as $3 per entry. No roster management, just draft and Rival sets your best lineup automatically. Playing fantasy baseball has never been easier! Sign Up Today!
What Clutch Pitchers Are Made Of
Strikeouts are the ideal outcome with runners on base, especially in scoring position. Contact can result in sacrifices (that move runners), or worse, base hits. Yes, a double play is a considerable help but again I would think you'd want to avoid any contact because there is too much room for error. If you're a pitcher allowing contact to set up a double play, etc., are you 'clutch' for getting out of that jam or is the defense to be praised for completing the play?
We'll delve into what 'clutch' stat is and how it's measured but, for or the moment, I will work with pitchers who best produce the 'ideal outcome'; the most strikeouts with RISP, or kRISP, since 2016.
So would Chris Archer come to mind when considering what was previously discussed? There are several in the chart who probably did but Archer isn't the obvious choice to top the list.
What I'm after here is if we can consider kRISP a key component of clutch; does one lend itself to the other and is it a reliable allusion? To justify using total kRISP in lieu of ratio, let me explain why it works better.
The above chart includes two additional metrics; total pitches and kRISP/Pitches ratio. We could assume the more strikeouts you have is a direct result of throwing more pitches; we could instead go by the ratio. If Archer's 1.73% is similar to a pitcher who threw half as many pitches, could we surmise they are equally effective? In a sense, yes but going by that logic envokes some basic statistical probability theories that can lead to a fallacy.; that we can expect, with a level of certainty, the less-active pitcher will maintain the ratio as his pitch frequency increases. In that case, I'll need to be more exclusive.
When running the queries, I ended up making the qualification no less than 5000 pitches. A sizeable sample has to be incorporated here; as we saw, the differences in kRISP ratio are fractions of a percent so there is no concrete way to differentiate them outside of using totals.
To elaborate, Archer, Chris Sale, Max Scherzer and Rick Porcello are within a hundred or so pitches of each other. Ranking the pitchers by kRISP ratio has Archer in front by .07%. Considering the standard deviation is 0.2%, it's not the best to use that stat as our factor for comparison.
Well then, could we say for sure that more kRISP is the direct result of throwing more pitches, thereby making a ratio comparison more legitimate?
The following data includes all 62 (starting) pitchers with at least 5000 pitches in the last two seasons.
With 41% of kRISP being the result of increased pitches, we have a low-moderate correlation and not the greatest relationship. We are now OK to use total kRISP as the basis comparisons and as a component of 'clutch' evaluation.
Now that we've found the most effective pitcher with RISP, does that also mean Archer is currently the best clutch pitcher in baseball?
We'll get to that, but first we have the league averages/ratios against what Archer accomplishes with RISP. For further context, Archer's events make up for 0.3% of the last two years of league RISP pitching data.
Most categories Archer is slightly better than league average. His strikeout to walk ratio is nearly three times better and there is a significant disparity in OPS. Some other bonus features of Archer include an 8.41 SO/W and a .538 OPS with two outs and RISP.
That last part sounds like more evidence to the cause of declaring Archer THE clutch pitcher. However, using those statistics in that manner doesn't give a lot of context, but using leverage (or clutch) does. Much like how on-base percentage simply tells you a batter got on, what OBP doesn't tell you is what wOBA can; a much more accurate representation of what can be expected when you come to bat. Maybe under those previously-mentioned conditions, Archer wasn't in a high-pressure situation; the team was way ahead or way behind. Looking over the last two years, Archer has seen 89 more batters when the team is behind; a 15% difference, so not exactly a slam dunk assumption.
Getting back to leverage, let's look into 2017. Whereas the average MLB pitcher sees a drop in performance when the game environment transitions from low leverage to high leverage, in 2017 the opposite was true for Archer. According to Baseball Reference's leverage evaluation, Archer got (in some cases marginally) better as situations got tighter during the game. Be aware that's not a regularity for Archer and alone doesn't firmly establish him as a high-leverage ace. That said, is Archer a true clutch pitcher because of his league-best kRISP?
While I'd venture to guess that managers generally want their pitchers at their best regardless of the team disposition, you ultimately want your pitcher to step up when the game needs to be kept under control to facilitate a tie or lead change.
FanGraphs measures Clutch and define it as "...how well a player performed in high leverage situations" and “…how much better or worse a player does in high leverage situations than he would have done in a context neutral environment.”
Well, how about a comparison of FanGraphs measurement of Clutch and kRISP? Are they mutually exclusive or can we find a strong relationship? First, I'll show the top 10 Clutch leaders for the last two seasons.
Just two names on this list (Lester and Cole) show up on the kRISP list. Looking at 18 different pitchers can't possibly give us a conclusion.
So then I took all 62 who qualified for the 16/17 top-10 kRISP pitchers and checked that against their two-year Clutch figures from FanGraphs. This ought to tell us if there is any relation and if we can say with little argument that Archer can be called one of, if not the, best clutch pitchers in baseball right now based upon the conclusion that striking out a batter with RISP is a/the defining quality of a clutch pitcher.
Here you'll see a pretty terrible...no, a completely insignificant relationship between the two. In fact, when you look at the three hexagons on the right in the negative clutch region, they are our top three pitchers in terms of kRISP.
So while getting a strikeout with runners threatening to score is a big accomplishment, I can now surmise that we can't point to kRISP as a major factor in determining the 'clutch' factor of a pitcher. It would seem that a big factor, in this case, is the inability to put complete context to the stat.
Archer's slider and Francisco Lindor might disagree.
