Can Statcast Tell Us What is Stranding the Mets?

This year,  the reigning National League Champions have faced a large number of unexpected difficulties. Certainly the potentially career threatening injuries to Matt Harvey, David Wright, and to a lesser extent (in severity, not importance) have each put a damper on the season. Every team suffers injuries, under performance, and the occasional off the field or clubhouse problem. They’re all part of the game. One thing, though, the one thing that is truly remarkable, that is driving Mets fans up the wall at the moment, is their batting average with Runners in Scoring Position.

Or rather, Runners in Stranding Position, as many Mets fans now begrudgingly refer to it. This season, the Mets are competing for the lowest batting average with RISP in MLB history. They are currently tied with the 1963 Washington Senators with the third worst average since 1914. Only the 1968 Mets and 1969 Padres are worse. The 1968 Mets were on the upswing, getting ready to have their magical 1969 Miracle Mets season. The 1969 Padres, though, they were one of the worst teams in MLB history, finishing with a 48-114 record.

Before I start, I must acknowledge that there are divided factions around these sort of stats.  Many argue it is a statistical noise, arbitrary sampling, or selection bias. I’m inclined to agree, slicing up the season into such small chunks and looking at the numbers is bound to introducemany problems in a lot of different ways. However, there is another part of me that is deeply curious about this situation, and perhaps some of you out there might be as well. Yes, looking at splits in this manner is very dangerous, it probably has no predictive value at all. Then again, we are talking about a team putting up historic numbers in this brand new Statcast era, and our fancy new tools are just begging to be used. So lets have a looksie and see if there is anything to learn.

 

The quickest and easiest way to apply Statcast data may be to look at the xStats numbers, since it does all of the work of sifting through launch angles, velocities, and figuring out the frequency of hard hit balls. It is an easy at a glance comparison. In these 900 PA, xStats suggest the Mets have gotten pretty unlucky. Their slash line ranges from 10 to 30 points below their expected line, but even with this boost they still have the worst xStats in MLB with RISP. Last in xAVG, xOBP and xOBA, second to last in xSLG, xBABIP and xBACON. They are the worst offensive team with Runners in Scoring Position regardless of whatever bad luck they may have faced.

Of course 900 plate appearances over the course of a season is a small number for a ball club, and these are taken from a very arbitrary sampling of data in the grand scheme of things. Then again, these at bats are also the most important at bats during the course of a game, and the xStats are suggesting there is more than bad luck going on here. It tells us the Mets are making the worst contact in major league baseball when runners are in scoring position, but it doesn’t tell us how or why. So, lets delve a little deeper into exactly what is going on with the ball once it leaves the bat.

For much of this analysis I will be using the following categories for launch angles.

LGB: Low ground balls, < 0 degrees, these tend to have a decent batting average, for a ground ball.

GB: Ground balls 0-10 degrees, these tend to be outs.

LD: Line drives, 10-14 degrees.

Ideal LD: Ideal line drives, 15-18 degrees, these have peak batting average when hit well.

HLD: High line drives, 19-25 degrees, riding the edge between line drives and fly balls, batting average drops rapidly with angle, but increased slugging makes up for it.

Ideal FB: Ideal fly balls, 26-29 degrees, these are where your home runs live.

FB: Fly balls, 30-35 degrees, still some decent power numbers but rapidly losing value with increased angle.

WFB: Weak Fly balls, 36-50 degrees, generally an out.

PU: Pop ups, 50+ degrees, nearly automatic outs.

The total distribution between BIP hit with and without RISP aren’t very different. You may have assumed this, but it is always nice to have data backing it up. With RISP the Mets are actually hitting more ideal line drives, and those have great batting averages, generally around .700 or higher when hit hard enough. They are also hitting more of the higher value ground balls (LGB) in RISP situations. If you looked at this alone, you would probably guess they might be hitting a tad better with RISP than otherwise, but we know that certainly is not the case. Comparing to the MLB wRISP numbers, the Mets are hitting more balls in the air, in general. More ideal fly balls, which is great for home runs, more general fly balls, and unfortunately also more weak fly balls and pop ups. They are certainly a fly ball hitting team with runners on base, where as the rest of the league tends to hit more ground balls with men on.

Here is where we start to see major problems with the Mets. Exit velocity is down across the board with RISP. Every category is down. Ideal line drives especially, and that is tragic, it was their one bright spot from the launch angle data. MLB seems to generally hit the ball a little softer when runners are in scoring position, with each category tending to drop a half mile per hour or so, but the Mets are dropping 1,2,3 mph in some of these areas, especially the highest valued areas like line drives. We know from past research how much a mile per hour rise in fly balls can do for home run rates, it can add about 15 feet to fly ball distances, perhaps more. The Mets are having a 1 mph drop in ideal fly ball launch angles. These are the average numbers, though, and average numbers don’t tell us much about individual batted balls. Perhaps they are hitting the ball very hard on a semi-consistent basis, while mixing in a lot of weak contact. The following table depicts the percent of balls in play hit above 95mph for each of the buckets.

This is a bit less bleak than the raw exit velocity figures. Yes, they are hitting the ball softer, on average, but the percent of very well struck balls isn’t dipping anywhere near as significantly. Compared to the rest of MLB, they are hitting the same percent of ideal line drives, the same percent of which are hard hit. The same goes for ideal fly balls. These are both positive signs. They are way under performing on ground balls, though. That may be the big separator between them and the rest of MLB, the number of hard hit ground balls.

How many of you guessed slightly weaker ground balls and a 1% drop in line drives is what drove the Mets to a historically low batting average with RISP? None of us did. I present these stats today not to illuminate a stark difference between the Mets and the rest of MLB, but to show you how small of a change is necessary to drop a major league club from average to historically bad. Or, at least historically bad in this one very specific area of the game. You can scream small sample size, and I totally agree with you. This is a small sample size, it is not predictive. It could change in a second. They could go 15 for their next 20 attempts and turn this all on its head. I’m not arguing otherwise.

I’m here to highlight how narrow the bands for success are in baseball. The affects of exit velocity and launch on the game aren’t necessarily linear. A very small dip in exit velocity can create a rather large drop in success rates. Hitting the ball 2 degrees further to the left or right can slice off chunks of your batting average. Batting a ball 1 degree higher can take 10 points off your BABIP. Or you can flip each of these three statements and say the opposite. This data is teetering on the edge between statistical noise and revelation. Who would have considered swapping out 2% of your lowest exit angle ground balls for your higher angle ground balls might result in dropping your batting average by 5-10 points? That isn’t a result I would have ever predicted. I may have guessed the opposite, to be honest. How many of us fans knew a 1mph increase in average exit velocity would result in a huge league wide surge in home runs? There were a few, I wasn’t one of them. This data serves to highlight just how much these very small differences can be magnified, and it may take us all a long time to build an intuitive understanding of how these numbers work and relate with one another.