Aroldis Chapman and the Cost of Risk Aversion

In order to keep their season alive, the Cubs had to win Game Six. They won Game Six. Tomorrow, they play for all the marbles, with one more win bringing the franchise their first championship in 108 years. From that perspective, tonight was a success. Full stop.

But that perspective is a particularly binary view of the world, with only good and bad outcomes, and no room for the shades of gray that make up real life. In this world, things can be somewhat good, or very good, or painfully awful, or just kind of not great. In this world, we have not two possible outcomes, but thousands of them, with differing levels of magnitude. And from a perspective that accounts for the different magnitudes of outcomes, this Cubs win isn’t quite as great as it could have been. This win came with a cost, and probably unnecessarily so.

Entering the bottom of the seventh inning, the Cubs led 7-2, giving them nine outs to keep the Tribe from scoring more than four runs. History tells us that teams in this situation go on to win 96.7% of the time. But that 3.3% is not a comfortable 3.3% when that 1-in-30 chance means your season ends. Joe Maddon wanted to drive that number down to 0%, so he had Aroldis Chapman getting loose in case Cleveland started to rally.

Mike Montgomery got Rajai Davis out to lead off the inning. Eight outs to go, win expectancy up to 97.6%. But then he walked Roberto Perez, putting it back to 96.5%, and Chapman started to throw. Carlos Santana flew out. Seven outs to go, win expectancy at 97.8%. Jason Kipnis singled: 96.9%. With Francisco Lindor and Mike Napoli due up next, Maddon didn’t want to entrust his second-best lefty with the responsibility of getting out the team’s two middle-of-the-order righties. So in came Chapman, with seven outs to go, protecting a five-run lead.

For Maddon, and all the other new-school managers this postseason, this has been the decision. Dave Roberts called on Kenley Jansen for a four-out save in Game Three of the NLCS despite his team leading 4-0, and left Jansen in to pitch the ninth even after the team pushed their lead to 6-0. Cody Allen pitched the ninth inning of Game One of the World Series even with his team up 6-0. Bullpens have been managed in a radically different way this postseason, but the strategy of doing whatever it takes to protect a lead, no matter how large, has survived the revolution.

The justification for these decisions is pretty simple: managers are erring on the side of risk-aversion, not wanting to see their inferior relievers allow a rally that forces their best relievers into a tough situation when they could just use the dominating guy first and avoid the potential stress altogether. If you put Pedro Strop or Carl Edwards or Justin Grimm on the mound in the seventh inning, and Lindor and Napoli keep the inning going, you’re going to have to use Chapman anyway, and now the game be wildly different. Slam the door, win today, and figure out tomorrow tomorrow.

But that’s the binary worldview again. That doesn’t allow for their to be different levels of useful wins, or for what happens today to have a real impact on what happens tomorrow. In the world we’re actually in, where shades exist, this was about as great of a loss as Cleveland could hope for, because the Cubs gave up their best kind of win in order to secure a slightly more certain lesser win.

For the Tribe, they lost, but they lost without throwing Andrew Miller or Cody Allen. They didn’t even throw Bryan Shaw. By getting down big early, Terry Francona was able to lose with his JV pitchers. And now, in the real winner-takes-all game, he’ll have his three best relief arms completely fresh and ready to go.

Joe Maddon could have had that, too. Had he resisted the urge to use Chapman to protect a five-run lead, he could have had the only reliever he really trusts right now completely rested for Game Seven. Instead, Chapman threw 20 pitches tonight, and warmed up three times — once to come in, and twice between innings — while recording five outs. Chapman will still definitely pitch tomorrow, but now, his outing will come with real uncertainty.

To be clear, we don’t know what pitching tonight will do to Chapman’s effectiveness or endurance tomorrow. No one does. It might not do anything. Perhaps the adrenaline provided from pitching in Game Seven of the World Series will overcome any kind of fatigue that might set in after throwing 62 pitches in the previous three days. Maybe Chapman, who has struggled with his command at times this postseason, will locate better at 99 than he would have at 102. Maybe the Cubs will crush Corey Kluber in the first inning and Chapman’s ability to work multiple innings will have no impact on the outcome whatsoever.

To pretend that this is an area where we can definitely show the precise cost being paid would be folly. But uncertainty in and of itself is a cost, and it’s one the Cubs will now pay, because they also don’t know what pitching tonight will do to Chapman’s performance tomorrow. They can hope the impact is zero. They can believe that Chapman will still be Chapman, throwing gas for as long as he’s asked to throw gas. But the ability to plan on Chapman entering the game in the seventh inning, and pitching you all the way through the ninth? You can’t do that anymore, because Chapman’s usage tonight made his ability to do it again tomorrow uncertain.

If they hadn’t used Chapman, the Cubs probably could have told him he was getting the last nine outs tomorrow, and prepared him to start the seventh inning, regardless of the score. That would have left only 18 outs for Kyle Hendricks and Jon Lester, the team’s two best starting pitchers this year, to tag-team their way through. But I don’t know how you plan to give Chapman nine outs now, and perhaps more importantly, I don’t know how you create a plan for tomorrow that doesn’t require a reasonable contingency if Chapman looks diminished when he enters. Now, you need a Plan B.

And so now, Maddon might have to choose which reliever he might trust to pick Chapman up if the workload proves to be an issue, and he’ll either have to pick from among the pitchers he didn’t trust with a five-run lead tonight, or perhaps one of the starters-available-out-of-the-pen that could have otherwise been used before Chapman took the mound. If Maddon had a two-days-rest Chapman tomorrow, there would be little harm in throwing both Jon Lester and John Lackey in relief of Kyle Hendricks, allowing him to remove Hendricks from the game early if need be.

Now, though, at least one of those two is probably hanging back in reserve. The Cubs pitching staff tomorrow could have been almost ideal, with the team capable of getting multiple innings from four really good pitchers, locking out all the pitchers that Maddon doesn’t trust from a game where Cleveland will almost certainly rely only on Corey Kluber, Andrew Miller, Cody Allen, and maybe Bryan Shaw.

But now, the Cubs will likely have to extend Hendricks longer tomorrow, as the bridge to Chapman is less secure, and the team can’t be as confident that Chapman will be the lockdown relief ace they acquired him to be. And the Cleveland blueprint is clear; if they have the lead after four innings, they’re going to make the Cubs score the tying run against Andrew Miller. Any hole Hendricks might dig in the third or fourth inning could end up being the Cubs grave.

None of this is certain, of course. Because it’s baseball, there’s a good chance that none of this will matter, and the outcome of tomorrow’s game will be determined by something completely unrelated to the Cubs reliever usage in Game Six.

But tonight, Joe Maddon chose to give into the fear of the 3%, believing that his secondary relievers weren’t capable of holding a five-run lead, and not believing in his offense to help them win the game back even if they did. He managed not to lose Game Six instead of managing to win Game Seven at a point in which Game Seven was pretty close to already secure. The emotional desire to not even flirt with blowing a lead, ending the team’s season in the process, overcame the rational reality that the game was mostly already over.

So Aroldis Chapman got five outs that the other Cubs relievers probably could have gotten. Those outs may well have been borrowed from Game Seven, in a situation where the Cubs might really be able to use a full-strength Chapman on the mound. But they won’t have that, because when push came to shove, Joe Maddon only trusts one of his relievers right now.

For his sake, I hope he doesn’t need that one reliever to bail him out tomorrow.

Dave is the Managing Editor of FanGraphs.

newest oldest most voted

You’re not even addressing one of the most important aspects of the analysis!

The fact that allowing Chapman to pitch didn’t change their win percentage much at all.

The 97% win expectancy went to maybe 97.2% using Chapman rather than one of their other relievers, especially facing 2 righties in the 7th inning in Napoli and Ramirez (no matter how good Chapman is overall, he’s still better v LHB than RHB and the Cubs’ other relievers, Edwards, Strop, Grimm and Rondon are better facing RHB).

You make is seem like using Chapman raised their win percentage from 97% to 100% or at least you’re not telling the readers how much it did raise their win percentage.

That’s a critical part of the analysis. If using him raised their WP by 1% or 2% then it very well could have been correct to use him tonight even with potential reduction in performance or longevity tomorrow. right? But you don’t talk about how much it did raise their WE in game 6 so how are we to know whether it was a good decision or a bad one?


It’s funny but I think Dave was thinking it would be obvious to us that this was a super low leverage situation. He was focusing more on the #1 theme of this postseason, which is the optimal usage of the pitching staff across games in a do-or-die series (with the benefit of more “off days” thrown in.)

That in and of itself is worth a whole article about how Maddon used Chapman last night.