Riley’s Believe It or Not: Production and Production

The 2021 season has been a slog for the Atlanta Braves, for all sorts of reasons. There are many reasons why teams expected to be good may end up with subpar records, but the Braves have been beset by more than a few of them. We won’t rehash them here. Instead, we’ll focus on a bright spot (well, sort of...): Austin Riley.

After a torrid six-week debut stretch that quickly cratered into some real poor hitting, Austin Riley came into the 2021 season as more of a question than an answer. He was handed the starter’s job at third base, but his career to date comprised an 87 wRC+ and 0.1 fWAR in 503 PAs. In some ways, that was an overly pessimistic outlook, as Riley had an above-average .333 xwOBA in 2020, which he underperformed to mirror his 2019 wOBA of .307. In any case, you all probably know that Riley has straight-up raked for much of the 2021 season.

For the year as a whole, Riley’s wRC+ sits at 136, and while he’s outhitting his xwOBA (.380 wOBA, .366 xwOBA), the latter is still quite good. He’s done a lot of things well, from pummeling the ball frequently (a barrel rate that’s nearly double league average) to drawing a lot of walks to avoiding grounders, and he’s victimized fastballs, including finally starting to hit even the harder ones quite well. The net result of all this good work is that among players with 130 or more PAs (chosen because it scales my preferred 200 PA threshold down by two-thirds, given that two-thirds of the season is complete to date), Riley is 44th in wRC+, 28th in wOBA (Truist Park is very offense-friendly these days), 40th in xwOBA (note that xwOBA is not scaled by park), and 21st in total hitting value generated. (Side note: Ronald Acuña Jr. hasn’t played for weeks and is still seventh in total hitting value.) While Riley’s success in 2021 to date is no guarantee of future offensive dominance, it’s safe to say that he’s gone from a speculative starter to probably having locked down a starting role for the foreseeable future, ascending into a spot in the Braves’ near-term core. In short — Austin Riley: he is now good at hitting baseballs.

So, Austin Riley’s been productive. Yet, he also hasn’t been productive. What do I mean? Well... here’s the thing: despite the gaudy offensive line, Riley somehow has negative WPA. (As a refresher, WPA is literally just a stat that tracks how a player’s hitting or pitching affect their team’s win probability of a given game. Each team starts with .500 WPA, and It’s also not like, “haha he has a slightly negative WPA” even though that too would be weird. It’s pretty negative! It currently sits at -0.24, and it’s been negative for about a month. (Riley was positive in WPA from May 16 through the first game of a June 21 doubleheader; he bounced around a bit after but was last positive on July 2.) Honestly, this whole post is pretty much just this chart:

I mean, on top of the Biblical-level tsuris the Braves have already dealt with, there’s also this? Major ugh.

While the season isn’t over, and there’s some time for things to reverse themselves (for both Riley, and well, the Braves’ season), I wanted to see how uncommon this sort of phenomenon was. To do this, I need to use OPS+ and the Baseball-Reference formulation of WPA, which is slightly different from wRC+ and Fangraphs, but the idea is still the same: Riley has a 130 OPS+ and -0.5 WPA this season. In baseball history, there have only been 18 players to have an OPS+ of 130 or higher, and a WPA of -0.5 or lower for a season. Of those 18, only five had more PAs than Riley already has this season. Only one such season has occurred in this millennium. This is a stupid club to be a part of.

By the way, if you make the criteria more like the Fangraphs values (136 OPS+, -0.24 WPA) you’re left with just Bill Mueller’s 2003 and Edd Roush’s very sad 1918 among players with 400 or more PAs in a season.

There are other ways to express just how ridiculous this has been, too.

Riley is that red bubble, obscured by the others. You might say, “So what? There are plenty of other bubbles in his vicinity.” But the bubbles here are scaled by number of PAs... what if we remove anyone with fewer than say, 260?

Okay, Riley is a bit more dramatic, but we’ve still got some guys with less exposure than him. Let’s push it up to a 360 PA cutoff.

Austin Riley is literally the biggest WPA underperformer based on his wRC+ among this group of 110 players. Fun. “Fun.” You can do the same thing with batting runs; this makes things slightly better for Riley, as he’s no longer the worst underperformer of “expected WPA” — but he’s still third-worst, and the only guy with 360 or more PAs “rewarded” for his above-average hitting value with a below-average WPA.

What about the other Braves? Is this some kind of team issue? No, the regulars all pretty much follow the pattern... except for Riley.

So, what’s going on, exactly? It’s not much of a mystery. The only real way you can have something like this happen is if you rake when WPA isn’t at stake, and do the opposite of rake when it is. To wit, Austin Riley’s 2021 splits:

Yep, that’ll do it. Rake in low leverage. Rake in medium leverage. Implode in high leverage. And yet... while Riley’s walk and strikeout rates take a nosedive in high leverage, it’s hard to say that all of it has been his fault. I mean, look at this:

I’m not a fan of using the Fangraphs “quality of contact” stats relative to what’s on Baseball Savant and available through Statcast, but still! In high leverage, Riley has a liner rate of almost 38 percent, yet his lowest BABIP across the three leverage types. He is not particularly pull-happy, and has the greatest rate of hard-hit balls. And yet, the results aren’t there.

Unfortunately, we can’t do these types of splits using Baseball Savant (at least not yet, hint hint to humans who are reading this who might be able to do something about that). I tried to get close to it by querying specific situations, but I don’t think it really worked out.

As you can see, none of these really mirror his sub-.230 wOBA in high leverage. In some cases the xwOBA is pretty bad, but in others it isn’t. All of these are extremely small cuts of his overall sample, though, so we’d expect a lot of erratic numbers here.

None of this really suggests it’s been a systematic struggle for Riley, but just a quirk of when certain results have occurred. We have “clutch” score exactly for such a reason, to compare performance in lower leverage versus higher leverage — a negative value means the player has performed better when leverage was lower; a positive value means the player has performed better when leverage was higher. In Riley’s debut season, he had a Clutch score of 1.11, which is really high. In 2020, it was -0.45. In 2021, it’s an absurd -1.81 so far — the second-worst in MLB. Who’s been more cursed than Riley? Former MVP Mookie Betts, with a Riley-esque 140 wRC+ but 0.05 WPA. But hey, at least Betts is in the positives. Riley, not so much.

As a final note, I want to highlight the insane stretch that we just witnessed that led to this. It’s hard to do the sheer madness of it justice. Riley played in four games, including a doubleheader, between July 26 and July 28. His batting line over that stretch: a cool .400/.438/1.000 with three homers and a 270 wRC+ in 16 PAs. His WPA over that stretch? -0.45.

• On July 26, in the first game of the doubleheader, Riley went 1-for-3 with a walk... but hit into a double play with the bases loaded and one out.
• On July 26, in the second game of the doubleheader, Riley went 2-for-3... but hit into a double play with men on the corners, one out, and the tying run on third base.
• On July 27, Riley went 2-for-5 with two homers. However, he hit into an out at the plate with the bases loaded and one out in the first. He finished the game with negative WPA.
• On July 28, he hit a game-tying homer... but down by a run with the tying run on third and one out, he struck out.

These aren’t the only ones, either. His WPA is substantially dinged by that crazy July 4 game against the Marlins, where he was thrown out trying to score the walkoff run at the plate after a would-be wild pitch, and also struck out to end the Braves’ ninth-inning rally. On May 21, he hit two bombs and finished with negative WPA. He had huge games on both April 21 and April 28, with a homer in each... and those two games netted out to just about zero WPA. Perhaps part of the story isn’t that Riley has gone ham when it hasn’t mattered, but that he hasn’t any “dinky single wins the game”-type occurrences. When Riley’s racked up WPA, he’s had big games. He’s had literally two games all season with a WPA over 0.06 and a game wRC+ below 100; Freddie Freeman has had six.

All in all, this isn’t a thing. It’s a quirk. We may not even notice this by the end of the season, especially if Riley gets a game-winning single after an 0-for-4 day or something a couple of times. Or, results may ladder themselves totally differently across leverage next year. The problem is not one of process, just of results, because if Riley wasn’t the league’s second-worst batter by clutch, the Braves would probably have a few more wins. And that could make all the difference. Instead, they’ll have to muddle through this as well. When a team is having a losing season, it’s always something. But in the Braves’ case, this is one of way too many somethings. Maybe they’ll overcome them. Maybe they won’t. I’m sure more weird stuff like this will keep happening anyway.