On Sat morning, I was watching the Mets fast forward from the previous night’s game against the Marlins. The following situation arose:

Top of the 9th inning, Mets on first (LoDuca) and Third (Carlos Gomez), none out. Score tied 1-1, Renyel Pinto pitching to Shawn Green. Announcers Gary Cohen and Keith Hernandez go practically ballistic on Marlins Manager Fredi Gonzalez for having SS Ramirez and 2B Uggla playing at double play depth.

Result: a one hopper to SS - Hanley Ramirez throws home in time to get Gomez, but the ball glances off of Miguel Olivo’s glove; everybody’s safe, one run in, Mets 2, Marlins 1. The Mets go on to put up a 5 spot in the 9th and win 6-2 (a solo shot by Miggy Cabrera off Wagner in a 32 pitch 9th accounted for the 2nd run).

That one hopper would have been the easiest double play ever: two steps to Hanley’s left and an unobstructed throw to first (Loduca’s probably not even 1/2 way to second when Hanley reaches the 2b bag).

This got me thinking, was Fredi showing an “independent” strategy by giving his players the apparent “choice” to opt for the double play in that situation. In other words, does holding the damage to one run (by conceding that run) give the Marlins the best chance to win in this situation?

Before you say: “no way Jose Reyes” (ouch, pretty corny), lets look at the numbers.

David Nichols has an Expected Runs table, which uses 10 years of actual MLB data to give you the expected runs any given situation will generate in the long run, and provides the probability that the situation will result in at least 1 run:

http://www.nssl.noaa.gov/users/brooks/public_html/feda/datasets/expectedruns.html

Runner on 1-3, 0 outs: 1.75 expected runs (er) and 87% at least 1 run (>0).

This means that more often than not, more than 1 run will be scored, and roughly 1 out of 10 times will the offense be shut out completely (as a life long Mets fan, it sure seems that the Mets get shut out more than 13% of the time in those situations!).

Lets start with the easy one:

The Marlins turn 2, and now:
0 on 2 out: 1.1 er (1 run in plus 0.1 er/0.067>1)

This means that by turning 2 in this situation, the Marlins have a 93.3% chance of getting to the dugout down only one run.

What are the odds of scoring one run in the 9th? According to D. Nichols ERT: 27.5%.

Looking at the infield in situation is more complicated:

If you play the infield in and you get the ground ball hit directly at your ss/2b, the runner on 3b tries to score, and is out at home (i.e. everybody on defense executes perfectly and the ump makes the right call), now you are looking at:

runners on 1st and 2nd with 1 out (0.92 er/.42 >0).

If you play the infield in and the runner “stays put” and you get the “safe” out at 1st:

r on 2-3 1 out (1.40 er/.67>0).

With none out, it seems more likely that the runner at third will be a bit more conservative with a drawn in infield, as there are still many ways to score with one out.

So,under the realistic, likely “best case” scenario, if the ss/2b gets the ball hit right at them, and the runner stays put, you’ve only gained 20 percentage points in avoiding the tiebreaking run, and there’s still a pretty good chance that that fatal second run will be scored.

Note also, that the 1.4 expected runs scored with 2-3 and 1 out is 0.3 er greater than the 1.1 er by getting the double play. This is why going for the double play early in games is a no brainer in most circumstances. (Throwing out the aggressive runner at the plate generates a residual er of 0.9, not much less than the 1.1 er of the dp).

What’s the downside of playing the infield in? More balls will get through the infield, and if so, the home team is now down by one with either 1-3, o out (2.75 er/87%>1 run deficit) or 1-2 0 out (2.5 er/64%>1 run deficit). If this happens, it’s disastrous to the home team’s chances of winning. See below.

A successful “infield in” play reduces the Mets‘ odds of scoring 1 run from 87% to either 67% or 42%. However, a “successful “look the runner back” infield in play still results in more expected runs, which means a better chance of scoring 2 runs.

Based upon the above numbers alone, it seems to me that taking the double play in this situation is not necessarily bad play, and could even increase your team’s chances of winning the game!

As for the “chance of winning”, a fascinating toy and managerial tool is located at

http://walkoffbalk.com/

This site uses up to 30 years of historical data to demonstrate a team’s chances of winning in any situation, any inning, of a ballgame.

With runners on 1-3, none out, walkoffbalk gives the Marlins a 14.3% (5wins out of 35 games - 1996-2006) to 20.8% chance (27/130 - 1977-2006) of winning.

Using the last 10 years, getting the DP results in:

none on, 2 out, one run in: 18.1% chance (10 years data) and 18.8% (30 years data).

Getting the double play could increase the home team’s chance of winning.

Let’s compare it with the others:

Successful “infield in”

2-3 1 out - 31.8% (30 yrs) 36.1% (10 yrs)
1-2 1 out - 42% (30 yrs) 45.7 (10 yrs)

Unsuccessful “infield in”

1-2 0 out, one run in: 10.3% (30 years) 13.3% (10 years)
1-3 0 out, one run in: 6.2% (30 years) 5.5% (10 years)

So a successful “infield in” roughly doubles the home team’s chance of winning, while an unsuccessful “infield in” halves it. Again, not so clear cut as to whether the DP is better than “infield in”.

The Variables:

Ahh, the variables:

The verdict will largely rest with the following variables:

• expected success rate of turning the double play
• expected outs from a play at the plate, and
• the number of balls that get through the infield that would otherwise turn into outs or double plays.

I believe that the numbers will bear out that it’s easier to turn two than it is to get a tag out at home, and clearly, more balls will get through a drawn in infield.

How many? That is the question for which I do not have an answer.

Other variables that I can think of include:

• Success rate of the closer in protecting a one run lead (i.e. is the closer better than the “average” which is based upon past results set forth in walkoffbalk?);
• Speed of runner on 3rd;
• Speed of batter and runner on first
• Batters due up for the home team in the bottom of the inning.

I’m sure there are others.

To those who advocate walking the bases loaded in this situation, walkoffbalk counsels otherwise, as it reduces the home team’s chance of winning to 15% from 20% in a 1-3 none out situation (30 year sample size).

I sincerely doubt that many managers would take the double play over an out at home to preserve the tie. Think of the headlines of the NY Post the next day if Willie or Joe Torre lost a game by conceding a run in the 9th inning.

My point is that based upon the above, I am confident that there are situations where it is an objectively “better”
play to take the “guaranteed” double play over a potentially tough play at the plate. A thinking manager, and fantasy General, will not dismiss any strategy out of hand; a deeper analysis will always yield a greater understanding of the game, even if it confirms that the initial “common sense” or conventional wisdom is correct, after all. That wisdom is what the true General should seek to contest.

I hope this post will provoke a meaningful discussion, and welcome others’ efforts to improve on this analysis, detail the flaws and identify additional variables which could alter the result.