Alex Gordon Barely Had a Chance

Imagine if, for some reason, you completely missed Game 7. Not only did you miss it — you didn’t hear anything about it, from friends or from family or from the Internet. You get home, and this is the first thing you see:

jirschele

What on earth has to be going through your mind? It requires special circumstances for a third-base coach to end up with a postgame interview. And why is this one smiling? He must’ve made one hell of a decision. You know what the rules are, with regard to attention paid to base coaches. They only get it when they’ve done something controversial.

People want there to be a controversy here. The way the World Series ended was final, conclusive. Salvador Perez, 100% absolutely, made the last out on a foul pop-up. There is no what-might-have-been with Perez’s at-bat. So many have turned to the play before, when Alex Gordon was stopped at third after sprinting on a single and an error. It’s a frantic search for closure that resembles a frantic avoidance of such, and without any doubt in my mind, if Gordon had been waved around, it would’ve made for an all-time moment regardless. But while we can’t say for sure that Gordon would’ve been toast, since the play never happened, it sure seems to me the odds were too strongly against him. Mike Jirschele did the smart thing, and Alex Gordon did the smart thing, and Salvador Perez did the following thing. Barring a miracle, sending Gordon would’ve just ended the game a few minutes sooner.

We can walk through some simple break-even math, using this win-expectancy spreadsheet I’ve had for years. All of the numbers are approximates, estimates, but they’re close enough, meaning the result is close enough. As Gordon arrived safely at third, the Royals’ win expectancy stood at about 15%. That’s going to serve as our, I don’t know, fulcrum? If Gordon rounds and makes an out, the Royals’ win expectancy drops to exactly 0%. However, if Gordon goes and is safe, the Royals’ win expectancy jumps to about 55%. So, in one direction, there’s a 15% loss, and in the other there’s a 40% gain. We’re left with a breakeven rate of about 28%.

You can quibble a few points in either direction, but that’s right in the ballpark. Maybe the rate’s a little lower, because Madison Bumgarner was pitching. Maybe the rate’s a little higher, because Salvador Perez was batting, with the platoon advantage, and he’s been considerably better against lefties in his career. Remember that he’s solely responsible for the only World Series run Bumgarner’s ever allowed. One needn’t argue the details; the breakeven rate was somewhere around, let’s say, 25-35%. That’s low! That requires a low frequency of success!

It’s not possible to sit here and conclusively determine Gordon’s odds of making it. All we can do is make an educated guess, but my guess is Gordon’s odds were truly much lower. Jirschele clearly agreed, based on his decision and based on his words. As he said to MLB.com:

“As soon as Crawford secured it, and I saw where Alex was, I just felt we had no chance of scoring him.”

From Jeff Passan’s column:

“Believe me, I wanted to send him,” Jirschele said. “I couldn’t do it. I didn’t want to go the whole offseason with Alex getting thrown out halfway to home plate right there.”

From Tyler Kepner’s column:

“Even though there’s two outs, if he’s going to be out by a mile, I’m not just going to give them that last out,” Jirschele said, adding later, “Once they got that ball to Crawford, I saw we had no shot.”

Crawford has a strong, accurate arm, Jirschele knew, and sending Gordon would have forced him to use it. But Jirschele also knew it would not take Crawford’s best effort to end the World Series. Crawford could have thrown 10 feet wide of the plate, Jirschele guessed, and Posey would still have had time for a tag.

“To me, right there, if we have a chance to score him and I feel it’s going to take a perfect throw to get him, I’ll send him and take a chance,” Jirschele said. “I just felt that there was no chance that he was going to score.”

You can’t accuse Jirschele of not thinking things through. He didn’t make his decision because he was afraid of the potential consequences — he made it because he believed it was, mathematically, the right decision to make. Coaches don’t think in terms of breakeven points, but then we don’t think in terms of win expectancy, yet we still have a good sense. Jirschele had a sense, and it was the correct one.

Consider the evidence. This is where Gordon was when Crawford first got the ball in his glove:

crawfordgordon1

This is where Gordon was when Crawford turned to face the plate:

crawfordgordon2

This is when the stop sign went up:

gordonstopsign

Now, one thing that’s true: as the sign went up, Gordon slowed down a bit. If he were being sent, he would’ve been a little ahead of his observed position. But he still would’ve been very close to third base. Crawford was in possession of the baseball, looking home, and he had his feet set. It would’ve been a race between Gordon’s legs and Crawford’s throw, and Gordon tops out around 19-20 miles per hour. Crawford’s throws don’t.

This year’s Fan Scouting Report has wrapped up. Among shortstops, Andrelton Simmons got the highest arm-strength rating. Brandon Crawford tied for third. He also had a strongly above-average accuracy rating. Where Crawford was standing on the relay, he was quite a bit further away than he usually is, but that throw home’s still well within Crawford’s comfort zone. His throw would’ve had to be okay, but it wouldn’t have had to be perfect.

I want to give you some sense of perspective. From the regular season, here’s a screenshot of Alex Gordon trying to score on a relay:

gordonindians

Gordon’s halfway home with the player throwing. A screenshot, taken moments later:

gordonindians2

The ball still beat Gordon home. Granted, this relay came from closer than Crawford’s would’ve, but Gordon also wasn’t nearly so far along, on Wednesday. Throws are faster than runners. Intentional balls are faster than runners. Residential speed limits are faster than runners.

You want math? Here’s math. Looking over highlights, I clocked Gordon at about 3.4 seconds from third to home on a contested play. Watching various inside-the-park home runs, I got times for those runners between 3.3 – 3.7 seconds. Watching highlights of Brandon Crawford relays, I got times of 2.2 – 2.5 seconds from Crawford’s hand to the catcher’s glove. Here, for example, I get about 2.2 seconds:

Crawford was about as far away there as he was in Game 7. Because his throw was perfect, he was able to nail a runner who was halfway down the line upon release. Gordon wasn’t down the line at all. So the throw could’ve been average, or even a little worse. And I think the odds of that, or some other mistake, were low. They were real! Crawford and Buster Posey don’t make that play 100% of the time. But maybe 95%, or 90%. Maybe 85%, or 80%. The breakeven rate’s around 70-75%. All Crawford had to do was throw the ball, and all Posey had to do was catch the ball, and they’ve thrown the ball and caught the ball before.

Perhaps things would’ve been different if Gordon had achieved his top speed sooner. Perhaps things would’ve been different if Gordon didn’t stumble ever so slightly coming around second base. But everything could’ve been different if it were different. Mike Jirschele could’ve waved Alex Gordon around. Definitely, it would’ve changed the box score. Probably, it wouldn’t have changed the result. The Royals almost didn’t lose, which means that they lost, and it wasn’t Mike Jirschele who caused it.





Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.

189 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Josh
9 years ago

This is fantastic stuff. Just clear as day. Thanks Jeff

(the other) Walter
9 years ago
Reply to  Josh

Absolutely clear — and I’d been thinking all morning “what if he’d tried to send him?” – I knew Fangraphs would have a good review of the play.
The one thing missing – and it’s strictly secondary since with what DID happen, no way Gordon makes it – is “how much time did Gordon lose in his running, getting to 3rd – the stumble at 2nd, and, was he going full speed? He didn’t LOOK it, but maybe he’s just really smooth – did anyone clock how long it took him to get to 2nd, then 3rd, this time vs other times/other runners?

jason
9 years ago
(the other) Walter
9 years ago
Reply to  jason

I’ll buy that it would have been exciting as hell if he’d tried — but I find the article here far more convincing than that one — the stillshots showing exactly where Gordon was when Crawford got the ball especially. I just don’t see him having had even a 10% chance of making it at that point…*maybe* if he hadn’t stumbled earlier, and maybe if he’d gotten to third faster overall…but not at that point…if he’d been thrown out at the plate, he’d have been crucified worse than Grady Little for leaving Pedro in too long. I still WOULD love to see a clocking of his time running to third, vs his (or someone else fast’s) best time to third, to see how much time he lost getting there…has someone done that yet?

JRM
9 years ago
Reply to  jason

Watching the play, I was sure Gordon would be sent, but then I saw he was about five steps short of where I expected him to be.

I wanted him to be sent, because drama.

I think Jeff’s analysis is stronger on every axis than Nate Silver’s (and Nate’s body of work is tremendous). The numbers come out roughly the same, but Jeff discusses it a little better to my eyes; more importantly, he uses the actual locations of the interested parties compellingly.

In the end, we still end with an estimate… but I think it was right for Gordon to stop. Too bad it went this way, but it went this way.

Now, as to Salvador’s decision to swing at the first pitch…

Andy
9 years ago
Reply to  jason

Jeff really doesn’t disagree with Nate’s numbers at all. Jeff just thinks that Gordon’s chances of scoring were much lower than the 30% Nate estimated would make sending him worth it. Nate himself admits he doesn’t know what the chances of scoring were, whereas Jeff makes I think a pretty good case that they were significantly lower than 30%.

pitnick
9 years ago
Reply to  jason

That whole article is predicated on the idea that “he only needed to have a 30% chance of success” and I think Jeff’s work here shows pretty conclusively that his odds weren’t that good.

ChadT
9 years ago
Reply to  jason

Nate is good with elections, bad with sports.

Mikemember
9 years ago
Reply to  jason

Nate’s bad with sports? What about that whole thing where he created PECOTA?

Richie
9 years ago

Yes, just click the link in the last paragraph.

Bill
9 years ago

I read Silver’s post first and I’m a big fan of Silver’s but, congrats Jeff, your analysis is better.

Silver: SuccessRate > .3
Sullivan: SuccessRate .3 Then SendRunner
Else
HoldRunner

Jeff has far more evidence supporting his opinion than does Silver, so I think I’ll agree with him.