Playing for the Tie

Last night in St. Pete, the Red Sox trailed the Rays, 4-3 in the top of the ninth.  Fernando Rodney opened the frame with six balls in seven pitches, walking Will Middlebrooks (who gave way to pinch runner Xander Bogaerts), and falling behind Jacoby Ellsbury, who then blooped a single into shallow left.

Read on for more and to weigh in on John Farrell’s strategy.

At this point, with two on and no outs, the Red Sox had a 44.1% chance to win the game (according to Tom Tango’s win probability charts).  Shane Victorino laid down a good bunt, moving the runners to second and third and increasing Boston’s win probability… to 44.8%.  Dustin Pedroia followed with an RBI groundout before Rodney fought back from 3-0 to strike out Mike Carp, taking a tie into the ninth inning.  You all know what happened in the bottom of the ninth.

Win probability charts, to the best of my knowledge, assume average players in all roles.  Victorino is an above-average hitter, as evidenced by his .351 regular-season OBP (and his .353 wOBA).  He’s 5-for-11 with two hit-by-pitches in the series so far.  Rodney is an above-average pitcher, as evidenced by his opponents’ .310 OBP against him in the regular season.  But Rodney couldn’t find the plate, perhaps opening a door for a big inning for the Red Sox.

Bunting was the right strategy to scrape one run across.  It put Pedroia in a position to bring the runner home with a groundout or a fly ball, and significantly reduced the likelihood of a double play.  But the Red Sox were playing on the road against a team that had won three straight elimination games and was flying high after taking advantage of Boston’s shaky defense in the eighth.  A big inning would have all but clinched a trip to the ALCS for the Red Sox.  A run just prolonged the game.

What should John Farrell and Victorino have done?  Was the bunt the right call?  Weigh in below.


Playing for the Tie — 13 Comments

  1. A bunt, when successfully implemented, raised the chances of WINNING the game. That’s not playing for the tie. The bunt raises the odds of scoring a single run in the inning which may seem as playing for the tie but in fact the winning run is only on first with no outs so moving him over to second with one out is also a somewhat reasonable approach to bringing him home as well. Classic small ball would have said have Elsbury bunt, not Victorino. That would have been playing for the tie.

    Now, the strategy of bunting is more complicated than win probability. Some guys bunt well, some guys don’t. Some guys swing well, some guys don’t. Victorino is a good bunter and quick down the first base line, particularly from that left handed batter’s box. If Ortiz is up instead of Victorino or something, that changes things pretty clearly away from bunting. Similarly, you’d be crazy to have your typical NL pitcher swing away in that situation. Factoring in the odds of losing a 1 run lead with Uehara, I think you are content with 2 runs, which is a very important point to our discussion!

    Who is Victorino as a player? He’s a free swinger. Don’t let his walk rate fool you, it’s 18 HBP to only 25 BB in 532 PA for a low 4.7% chance of a BB. Rodney’s missing the plate but he’s now 7 pitches in and likely a lot looser. I would not bank on a walk and asking Victorino to take pitches is likely a bad idea as well, it’s not his game. Victorino’s pretty good at avoiding an out, but he’s got good hitters (pedroia, carp, napoli) up behind him as well. Again, Victorino has good bunting and speed and with his noted issues with avoiding an out, I think bunting is the right call. Victorino may even beat it out.

    To me, the decision is just as much about getting the winning run home from first as it is getting the tying run home from second. It’s easy to look at it as hated smallball “move a guy from second to third with no outs” but that’s neglecting smallball 101 “move the winning run from first to second with no outs”. Elsbury is moved out of the double play and into scoring position as the winning run, not the tying run.

    Now, the question to me becomes what was Pedrioa told to do. I would have told him to avoid an out, not sacrifice himself. If that ball gets by SS, a hustling Elsbury probably gives Uehara a save situation. Assuming that was the case, I have no problem with the manager decision to tell a good bunter to bunt in the most classic of smallball situations.

    • Good point about Victorino’s walk rate. I think Victorino’s speed is key to both sides of the argument. If there’s a decent chance he bunts for a hit, WP goes way up with bases loaded and no outs. For all the walks he gives up, Rodney is tough to hit, so a bunt may be a reasonable way to get on base in this situation.

      On the other hand, as many have mentioned, a faster runner is less likely to hit into a double play, and avoiding the double play has to be as important to the bunt/swing away decision as moving the go-ahead run into scoring position.

      And I think Pedroia was trying to hit something hard and settled for an RBI groundout. If he were really trying to sacrifice himself, he probably would have tried to go the other way, moving the second runner to third as well. I’m sure that’s not as easy as it sounds, though.

  2. Bryan:

    With Victorino at the plate, the bunt has to be the better strategy, not only for the reasons mentioned by mosc above, but also because, despite his performance in games one and two of the series, Victorino was a guy who batted .161 in Tampa during the season, meaning that to expect another hit at this juncture would be to rely on gut feeling, not statistical probability. He’s a good bunter. Let him make one of those ever-popular productive outs.

    • Victorino’s .161 BA in Tampa was over 31 AB. That’s pretty much the definition of small sample size, so it’s hard for me to see that as particularly meaningful/predictive.

      • As I suggested, that is just one element of several, the others delineated in mosc’s contribution above. This isn’t to say Victorino might not have gotten a hit of some kind or worked the pitcher for a walk. The question, though, was to validate or question the strategy of the bunt in that set of circumstances.

        To me it wasn’t a hard decision to have him lay one down. It was the best choice.

      • Yeah, but we’re trying to make a meaningful decision or predict what a single player will do, so it’s more relevant/appropriate than just using WPA, I feel. If there’s one thing I’ve learned from teaching stats, it’s that they are more useful numbers than many people give them credit for, but they simply provide some context for you to make an informed decision.

        Not that I’m saying I believe baseball managers knowingly use stats wisely, but in this situation, I see the context that mosc laid down as an appropriate set of considerations for a decision of a bunt. Maybe we can dig deeper with numbers and find whether Victorino’s bunt success rate is greater than league average, or whether his strikeout/gidp rate is higher than average (the 2 outcomes they would want to avoid to not even be playing for the tie anymore).

  3. Surprising that 1st and 2nd and no outs is essentially a wash with 2nd and 3rd and one out, when trailing by a run in the 9th on the road. But, if the hitter is not that fast and hits a lot of ground balls, I would definitely be bunting.

    But, I’m going to say the bunt wasn’t a great move here simply because Victorino is a savvy hitter who will be selective with a pitcher who is not hitting his spots, and because his speed makes him an unlikely double play candidate (yes, I know he’s banged up but he was still plenty fast tonight beating out the slow roller in the 7th – it wasn’t even close).

  4. mosc: “Now, the strategy of bunting is more complicated than win probability”

    Bingo X 10.

    At the risk of becoming an absolute broken record on the bigger issue involved in these kinds of questions, and to keep the story short, using Win Expectancies to predict the usefulness of a move over some large sample of games of randomly chosen teams over some randomly chosen time frames is one thing…

    …and using them to predict what Victorino, Pedroia and Carp will accomplish against Fernando Rodney in Tampa in 2013 is **another thing altogether**. Managers are paid to make exactly this type of *context specific* decision, day by day, pitch by pitch.

    Is this fact really that hard to understand?

    The major negatives are just standard baseball thinking: (1) Victorino is fast and a lefty and so presumably a low DP candidate, and (2) they might well walk Pedroia to re-establish the DP possibility (and get to Carp) anyway. If they don’t do the latter, then you’ve got Pedroia up with guys on 2nd and 3rd with 1 out and no DP possibility, and I’ll take that any day. If they do, then you’ve got other options, including the suicide squeeze, which at least gives you one run. And remember, the Red Sox are up 2 games to 0 here. If the strategy fails and they lose, they’ve got two more chances to win a game, including one at home.

    So, I’d do what Farrell did any day. The real problem is what happened in the bottom of the ninth.

    • “Is this fact really that hard to understand?” No. Are you suggesting that someone doesn’t understand this, Jim?

      I considered the possibility of walking Pedroia as well, though Carp is a lefty and hit well this season, and with Rodney struggling to find the plate, I don’t think loading the bases was an attractive option for Maddon.

      One more note: Victorino has been a switch hitter for much of his career, but he stopped hitting lefty after breaking his thumb midseason. His speed certainly affects the likelihood of a double play, but his handedness doesn’t.

  5. “Is this fact really that hard to understand?” No. Are you suggesting that someone doesn’t understand this, Jim?

    No, I’m not suggesting that you don’t understand it Brian, I’m really not. I know you’re knowledgeable on these things. I want to make that absolutely clear and I mean it.

    My question there was an expression of a general exasperation of seeing people attempt to apply statistics that were either (1) generated from some (typically, vastly) larger collection of situations, and/or (2) really only applicable to such situations…and then applying them to very context specific situations. I just see it over and over again (not necessarily at HHS, but more generally). I should have been more clear on that.

    Good point on Victorino no longer batting from the left side, thanks for that (and it does make a difference, relative to DP potential).

    • Thanks for the clarification, Jim. A manager has to be armed with quantitative and qualitative data to do his job. Making decisions based solely on WPA, as you and others suggested, ignores human factors like who’s feeling 100% healthy and who feels comfortable in certain situations.

      Conversely, making decisions without some knowledge of generic WPA data can lead to awful decisions like the ones Fredi Gonzalez made in the NLDS- bunting just because your predecessor and his predecessor would have bunted in this situation, sometimes based on observations from deadball-era baseball, when one run won a lot of games.

      I used WPA, at a stat-based website, as one of the statistics I thought was relevant to the conversation, which I then opened up to readers to elaborate on other relevant factors. I’m sure you understand this, and I’m not accusing you any more than you were accusing me, but it’s not uncommon for commenters in this forum (which really is the best forum of its kind) to come across as hypercritical of posts that discuss an issue without addressing it from a certain angle. We’re just starting conversations.

      I apologize if I come off as oversensitive to such criticism. Back to the playoffs. They’ve been phenomenal so far.

      • No you didn’t come across that way at all–and I can see how my wording could have come across as overly critical. Indeed, it WAS critical, but that criticism was meant to be general. And I’m also aware that it’s Bryan, not Brian; I really need to do a little more proofreading every now and then…

        Anyway, an extremely interesting question (to me anyway) is the optimal weighting and integration, in a manager’s mind, of the general, versus context-specific, information that he (potentially) has at his disposal, in the making of any particular decision. But then, I’m much a geek about such things.

  6. Normally, I’d frown on giving away an out against the likes of Fernando Rodney, who has been shaky all year. But given Victorino’s superb record of bunting, and Pedroia’s equally outstanding record of getting the man home from 3rd in that situation, I liked the bunt.

    Pedroia’s K rate is under 9%. His career rate of producing that runner from 3rd with less than 2 outs is 60% (league average is 51%), with just a 28% failure rate (12% were BBs or HBPs, no run but no out).

    As others have noted, Farrell wasn’t strictly playing for the tie — who wouldn’t like having Pedroia up with a chance to drive in the go-ahead runs with a single? But even so, I felt the BoSox held the advantage in a tie game at that point, as they had not yet used Uehara, while the Rays had used all their best relievers. That angle didn’t work out so well, but I still believe in it.

Leave a Reply

Your email address will not be published. Required fields are marked *