Wednesday, October 07, 2009

Bottom 7 (when losing is your optimal strategy)

This past weekend in the finals of the mixed division at South Regionals, Bucket jumped out to an early three break lead. Pulling at 7-4 I turn to a teammate and say,"if we score this point or the next one, Jukebox will start subbing deeper so they can rest their studs for the backdoor final." My teammate tells me I'm crazy, a team's got to play all out in a game to go . Jukebox ends up scoring that point and getting a break back on the following one to bring it to 7-6 before we take half 8-6, receiving to start the second half. Jukebox continues to sub tightly through the remainder of the game as we trade out to win 13-10. Jukebox then goes on to face Rival (in a game that has been written about elsewhere), a team they had beaten the previous day 11-4 and that on paper does not match up with them. At 6-6 in that backdoor final my teammate turns to me and says "maybe you were right." Jukebox ends up pulling the game out 15-14 (in controversial fashion) but I believe they could have locked up that second bid much easier by resting their top players in the front door game.

In most tournaments it's rarely a question if winning is your best strategy. But in some formats, particularly those where multiple teams advance to the next round are there times when losing is your best strategy? There are some obvious downsides to pursuing a losing strategy - the obvious one is that in most cases you are giving up an opportunity. In the Jukebox situation, sitting their studs in the second half means giving up the opportunity to take the front door bid and pinning all of their hopes on winning the backdoor one. Maybe a bigger downside than reducing your opportunity, there is a mental challenge to admitting defeat. How does it affect you and your team's mental game to at some point say "we have no chance in this game, let's rest up for the next one"? This can be discouraging to some, and it may be impossible to convince some teammates that losing is the best strategy. Finally, no one will ever fault a leader for trying to win. You try to win the front door and fail and then fail in the backdoor because you put your efforts in the front door and most people understand. But if you give up in the front door to put your efforts into the backdoor and then go on to lose the backdoor? ...well, I'll bet there'd be hell to pay for that one.

I don't have the games in front of me, but there were multiple situations in the late '90s in the Atlantic Coast college open division where the loser of the front door game to go would go on to lose in the backdoor game to go. Given UPA regional formats this is a frequent occurrence in two bid regions where the top three teams are around the same level and then there is a drop off. The front door finalists battle tooth and nail, while the back door finalist cleans up waiting for the front door loser to finish exhausting themselves in a tight finals game.

Another situation where I've experienced something similar is in the Club Nationals format. Club Nationals is a brutally long tournament where there are a handful of games that have little to no meaning beyond keeping a teams confidence up. As long as you are able to make quarters without going through the play-in game your chances are pretty much the same. In 2005, Bravo faced Jam in the final power pool game on Friday. The game was to determine who was the 2nd overall seed and who was the 4th overall seed going into the quarterfinals on Saturday. The winner had a marginally easier road in elimination play - either way though both teams would have to beat 3 good teams to win a title. The game was hard fought with Bravo subbing tightly throughout most of the second half. After we turned over three game point possessions Jam was able to punch in the goal on a sort-of greatest to win 17-16. Bravo was crushed. We had completely invested emotionally and physically. After the game finished the team acted like it had been completely eliminated from the tournament. Even though the champions the three previous years had each lost at least one game going in to bracket play, it felt like our season was over. At that point, playing DoG in the quarters and losing (15-8ish) felt like a simple formality. My take was that the investment in Thursday's Jam game had drained our ability to play. This probably was not a situation where we should have intentionally tried to lose, but in my estimation it was one in which we should have not put such weight on the game, subbed a little more openly, and kept better perspective of the long run goal. Others on the team have said that winning would have given us an easier game against Pike in the quarters and the mental boost from winning the tight game, winning the power pool, and beating a team we had frequently struggled against would have pushed our confidence and game to where we needed to be to make a run at the title. How should teams deal with games like this, both subbing wise and mentally?

In some situations, there is literally no format advantage gained from winning a game. In a four team pool if the 0-2 team plays the 2-0 team in the final round of pool play, the result of the game has no impact on the team's tournament match-ups going forward. In some situations, a team locks their pool play position up with point differential before winning a game such that points moving forward in that game have no practical impact on the teams chances in the tournament. One interesting situation like this occurred at 2008 UPA College Nationals. Pool D (as is often the case in this format where the 4, 5, 9, and 16 are in a pool together) was a mess going into the final game of pool play. The seeding stood like this:

D1. Michigan (1-1) (Loss to Georgia 13-15, Win against Harvard 15-12)
D2. Texas (0-2) (Loss to Harvard 16-17, Loss to Georgia 13-15)
D3. Georgia (3-0) (Win against Harvard 16-14)
D4. Harvard (1-2)

As it sat with the Michigan v. Texas game still to be played, Georgia had won the pool. In order to clinch second in the pool, Michigan had to beat Texas or lose by 1 point. Texas had to win by a point to clinch third in the pool, and two or more to clinch second. For Texas, at any point in the game the strategy was clear: win by as much as possible.

This game was an all out brawl with both teams going point for point. Harvard stood on the sideline, their day finished, knowing full well how the point differential worked out. They needed Michigan to win the game. At 13-13 Michigan held on offense to take the lead 14-13 and wrap-up the second seed. At this point, Michigan could gain nothing from fighting on. Like all of the teams in the pool they'd had a brutal day with all of the games being decided by three or fewer points. Michigan's top players including Will Neff, Dave Fumo, Ryan Purcell, and Ollie Honderd, were starting to show signs of wear only overcome by their competitive drive to win the game. Texas of course, had to continue to push. Without a win they'd be relegated to consolation, a bitter pill to end a promising season.

Michigan, either unaware of the situation, having decided that winning trumped the effort they would need to put in to win, or feeling like it was their duty to Harvard to play as hard as they could, kept their rotation at around 10 players. Neff, Fumo, Purcell and Honderd played almost every one of the remaining six points as the game dragged on to the final score of 17-16, with Michigan winning. And those six points looked like the hardest of the day for Michigan, long points with lots of turns. At one point Fumo got a goal saving lay-out block and landed hard on his hip. Neff was clearly struggling between points only to put his all in to each meaningless point.

Well, meaningless for Michigan. Michigan's win sent the Harvard team, watching on the sideline, into pure joy as they rode their own one point win over Texas in the first round to Saturday elimination play at Nationals falling to eventual champions Wisconsin in the quarters. Texas meanwhile, with 3 losses by a total of 4 points, wound up battling in the 9th place bracket. For Michigan, the fight at the end of the game seemed to have taken a real toll as they lost to Illinois, a team that Michigan had handled 15-11 in the finals of regionals, the following morning in the pre-quarter round.

I apologize for the length of this post. I thought it would be a good discussion following Martin's earlier one about when to play your top 7 and I've been wanting to write about that Michigan/Texas game for a while now. But I'm interested to know others' opinions on coaching/leading in situations like these. Have you ever found yourself in a winnable game but chose to pursue a losing strategy intentionally because of larger goals? What would you do in these types of scenarios? As a coach or captain, how did you communicate with your team so that they understood the choices being made?


AJ said...

Nice post Kyle:

I'm a big believer in playing the format. I can think of three times (worth mentioning) where a team I have coached or been on has "taken the loss" in order to maximize our chances.

2006 AC Women's Regionals--I was coaching the Emory women...I talked about it in detail here: Basically, we had to win 3 of 4 games and it didn't matter what order we won them we played open rotations until we took a loss.

2008 AC Regionals: I was coaching Georgia and there were two bids to nationals available. We came in as the three seed behind Florida and UNC (which had been the finals match-up at Centex). We won a hard game in the semis against UNC and then had Florida in the finals on Sunday morning. I think with those teams we would beat Florida maybe 2 or 3 times out of, coming into the game my strategy is to keep the lines wide open play everyone...if we keep it close go for the win late, if we don't go with all third line guys. We lost 17-4or5. Embarrassing for sure, but we were able to beat UNC in the backdoor 15-13.

2007 Club Nationals Power Pools. Chain is 4-0 Sockeye is 4-0, we're playing for the 1 seed overall. I didn't see any advantage to winning the we rolled over, losing 15-4ish. When we talked about rolling over, we knew that we were going to wish we had played a harder game if we lost the next day in the quarters.

In any event, I personally was as fresh as I've ever been on Saturday of Nationals and by losing we set up a quarters match-up with Goat, a team I had never lost to, and then a probable semis against Bravo, who we had beaten the previous 4 times we played them....We go down early against GOAT and aren't able to bring it back in time losing 15-14. Insanely frustrating...I'm honestly not sure it was the right call. We certainly didn't look as sharp on saturday as we had looked on thursday and friday...maybe some of that had to do with packing it in aganist Sockeye...on the other hand we were really really fresh on saturday...given the same circumstances I would make the same call I made then...but who knows.

parinella said...

Good topic and I'm not sure of the answer, either. I could pull out anecdotes that support any viewpoint you like.

Michigan didn't lock up their position until they scored their 16th goal. Until that point, they could still have lost 16-14 or 17-15 and taken 3rd with a presumably tougher pre-quarters.

I think there is a big difference if you have an overnight rest instead of 30 minutes or even a bye off. In almost all cases, if you have an overnight rest, I'd play to win. With the 30 minute rest, maybe not.

A related question is when you're clearly better than the opponents (say, a seven goal favorite if both teams played all out). Too many times, we've been in that situation and played studs less and (more importantly) everyone played a bit flat, and before we knew it we'd be in a tight (and draining) game late. In this case, I think the proper thing to do is to make sure everyone is fired up and playing hard. What you do with subbing here is less important. Chances are your studs are going to play some, so that means they have to get warmed up and will do some sprinting anyway. I won't say that this is how we've played these games, but it seems to me today that it's the right strategy.

And on another related note, that 7-4 point you mentioned is the most critical one for the leading team. Score that, and the other team might pack it in in anticipation of their next game and you don't run the risk of losing in heartbreaking fashion.

It'd be fairly simple albeit time-consuming to go through all the Regionals to look for games where the final game-to-go is a rematch. Analyzing the data might be tricky, since a slight favorite would only win 55-60% of the games anyway.

mark said...

kyle - if i recall, a similar thing happened to Michigan playing UNC 4 or 5 years ago at the college championships. UNC won the meaningless point (for them) and sent Michigan to the consolation bracket.

When i was at Duke in 2002, we had a similar quandary in the frontdoor regionals game. We played in the finals against UNCW, and prior to the game the captains told us that we were putting it all out there so we could finally get the UNCW monkey off our backs (we very rarely beat them from 2000-2002). We got down a break early and basically traded the rest of the game, to lose 15-12. It was 95 degrees and we played maybe 12-13 deep, so we were pretty gassed after the game. W&M came out strong in the backdoor and completely crushed us - they'd had a comparatively easier win over NCSU at something like 15-8.

The thing is, I'm not sure we could have mentally worked through "giving up" in the finals (and W&M might have just been better, too), just as I'm not sure that Michigan/UNC/Bravo could have turned the switch off at such points. Some of it depends on your team and its attitude, but as AJ's comment indicates, even a fairly "easy-going" team (that's fair, right?) like Chain had trouble finding the same focus after relaxing for a game.

My guess is that it's more a psychological problem than a physical one - if there's a way to trick your team into staying sharp while not completely committing itself to a meaningless game, you can lose that game and not be affected by it. I'd also guess it's harder to do at the club level than college; college teams typically rely on a few studs to get things done, and if they all understand that they need a game off to recover, it's easier to give them a rest and fire them back up when it really matters. If you assume a club team has 10,14, even 20 guys who need to be sharp to win the big games, then it's much harder to have them all buy in and get the mental focus back after losing a game they may have been able to win. More importantly, that club team would have spent a game where those critical guys would have "practiced" playing at less than 100%, which may be the biggest problem.

Tarr said...

In neither of the situations you discuss is there an incentive to lose. There's essentially never an incentive to lose.

What there is, what exists in ALL formats with multiple games, is an incentive to save your legs for later games. As such, expending the energy of your best players in a situation where the chance that effort will help you is low (most obviously, in a game you don't think you're going to win, or a game where you've already built a large lead) is not a good idea.

As AJ says, this counts as "playing the format", but it's misleading to say there's an incentive to lose. There's not.

BenjamminSpears said...

kyle, interesting discussion. you give some great examples of why to play a bottom 7 and when to think twice about it. also like reading aj's talk about coaching - then playing - through some of these dilemmas.

[[having captained and coached teams that were constantly fighting to reach the top tier (battling in the back door of tournaments like regionals and nationals), i don't have many similar examples.]]

lookin forward to more posts like this one. congrats on qualifying for natties!

Kyle Weisbrod said...

Jim, you're spot on about Michigan needing to win by two and not having the spot locked up until 16s. I can't believe that has been bothering me so much for 18 months and I was just wrong. Weird. That said, they had clinched their pre-quarters spot once they scored their 10th point. But yeah, that's a lot more tricky situation than I made it sound in the post. Those end points weren't meaningless.

Tarr, I think you're getting hung up on the "incentive to lose". I don't believe I ever said that. There are, however, some situation where there may be an incentive to lose by greater rather than by fewer points. For example, at Club Nationals where losing by a greater number of points in a three way tie in your day one pool brings the team that you beat down to the lower pools allowing you to go in to the second day 1-0 as opposed to 0-1.

In an e-mail to me Sam Rosenthal brought up a 2002 Epig situation at Club Nationals where this exact situation played out. Boss Hogg, once the game was unlikely to be won, tanked a game against Machine so that Epig would drop to the lower pools with BH.

Tarr said...


The parenthetical title of the post is "when losing is your optimal strategy". The answer is, "never". However, playing your bottom 7 can be your optimal strategy. I don't debate any of the substantive conclusions you're drawing, only the language you're using to talk about them. Maybe I'm quibbling over meaningless details.

I'm pretty sure Day 1 of the 16-team club nationals format presents the only possible scenario where losing by more is better than losing by less. (Note: that format isn't in the manual.)

Tarr said...

I take it back; the pool-play completion pools in the 9 and 11 team formats also can present these perverse incentives, where if you are losing you want to lose by more. These are sectionals-only formats, but it can happen.

parinella said...

That's a good point about there always being an incentive to save legs. Baseball managers have this too with giving off-days and pulling pitchers earlier than they absolutely have to. I guess it's a matter of degree then.

Most of the time, you are dealing with small changes, decimal points in your percentages. We'll win 95% of the time subbing to win, 94% of the time if we rest our studs, so in the semis, we'll have a 51% chance of winning instead of a 50%.

In these cases, we are talking about large swings in each game, and (what galls people) a greater chance that the lesser team will advance because they are playing a tired team. The team in waiting can go from underdog to favorite against either team, provided that the first game to go is a battle. (This differs from a similar situation in the front door because the team in waiting earned that advantage by dispatching their opponents more easily, not by losing an earlier game and having an easier schedule as a result.)

Nathan said...

I have intentionally lost a game - in 2000 or so, the Spice Girls intentionally lost a game in pool play at WUFF Coed Nationals.

As I recall, we had been crushed by Ken Dobyns and Mike Gerics and others in a pool play game. A couple other "upsets" happened, and it ended up that if we won our last game, we would be on the same side of the bracket as them on Sunday, while if we lost, we would be on the opposite side (the details escape me). We perceived the other pool to be really weak, so we didn't really care which team we would face in quarters - we just wanted to stay away from those guys (nothing personal...they just beat us badly).

So we threw the game - I discussed with a few others on the team, then presented the idea to the rest of the team. Definitely there was some grumbling, but we did it. Actually, I think we lost at double-game point, receiving, which meant we had to sub in people we were sure were committed to the plan on the last point. We tried to disguise the turnover as just a poorly executed huck and then stayed 2 steps behind them on the turn.

And, well, it worked. We cruised through to the finals (beating some old Boston guys along the way, watching some other Boston guys collapse after being up 12-6, game to 13), the team we were afraid of lost somewhere along the way, and we beat Hummer and company handily in the finals.

I don't know what that says about anything, but it is an interesting story.

Tarr said...

Jim, is it really the large swings that make it seem galling? I don't think that's it. I don't even think the rematch factor is needed to make it seem galling. Basically, any format except pure single elimination is going to have some "oh, it wasn't worth winning that game" scenarios.

Let me throw out another example. Last weekend, Operation Kapow played Cougars in the quarterfinals on Sunday. This was the pool-based 16 team, 3 advance, where the semis are games to go and there is no 2/3 backdoor game. Kapow had beaten Cougars handily at sectionals.

Kapow wins the game, loses a tough game-to-go against LA Metro, and then loses another slugfest to the other top Socal team (Brown Chicken Brown Cow) to get eliminated. Cougars, by contrast, got a relatively weak Hawaii team in their first elimination game, followed by the Fort Collins team (who had beaten Brown Chicken Brown Cow in the first round, in a game most people considered an upset). In the 3rd place game, Brown Chicken Brown Cow was worn out by their Kapow game, and Cougars end up taking 3rd.

In this case, you could make a strong argument (and several Kapow players did) that Kapow would have been better off losing the first game, and winning two elimination games against relatively weaker teams, rather than play two tough games against the SoCal teams. Now, in the abstract this is almost certainly wrong. If you really think you're better off losing that game, then you should win it, and throw the next game, against LA Metro. Then you're fresh against BCBC, and you probably win.

This gets to the fundamental issue in nearly all of the cases where people say "we would have been better off losing". These teams do a poor job evaluating their chance of winning each game, and subbing to keep their top lines fresh for important later games. The problem isn't that these teams win the games they win, it's that they fail to sub down the roster when it's likely more hard games await them.

Most teams are a poor judge of when they are better off opening up the rotation to be fresh for later games, and even when they judge this to be the case, most teams do a poor job executing that plan. I myself have made the mistake of holding out for a sliver of a chance to win and tiring out top players, so I'm not saying this is easy.

So, the "should have lost" complaints will persist unless we either go to single-elimination formats, or only play one or two games per weekend.

Kyle Weisbrod said...

Thank you, Tarr. This is what my post was supposed to be about. It is not a critique of the formats. Like AJ's question to Martin in the game-to-go on Sunday, "When do you put out your top 7?" I'm asking folks, "when do they put out their bottom 7?" Or rather, when do they rest the studs and open up the rotation?

How close do you need to be to keep fighting? If you make the decision to stop fighting, how do you prevent the potential emotional or mental letdown? What is the mental impact of a win in a game that doesn't have serious consequences on your team's chances (like the power pool game for the top two teams in the pool at Club Nationals)?

Andy said...

hmmm... interesting topic, but I have an unrelated comment.

I was wondering if you would do me a big favor and tell me which ultimate videos available online are your favorites. I am trying to find some videos to inspire/instruct new and intermediate players on my team. Most of the videos on youtube are mostly highlights, and the instructional ones are rather old. I would appreciate any of your expertise on this subject.


Martin said...

Buy the Worlds video from 2008. It has the best filming and commentary that I recall. Also the UPA Nationals video from the same year is pretty decent. Commentary is pretty good, although it sounds like they edited out comments at times (Kyle, do you remember if that happened) and I didn't like their video editing choices in general.

Back to Kyle's comment: It is so hard to answer those questions because each team is so different. For a young team, I think the win is pretty important to build confidence. Although Bravo isn't a young team, your story would suggest that the emotion of that win would have really helped on Saturday. For a veteran team like Jim's (especially towards the end of their run . . . or now) it is probably easier to flip that switch.

Either way, these seem like great coaching moments. It is easier to adopt the "save your legs" strategy if it comes from a coach than from a captain, and it is up to the coach to be able to make that decision based on how their team functions. In 2006, coaching with AJ, it felt like we were playing for 3rd the whole weekend. But (most of) the girls bought in and we accomplished exactly what we were looking for.

I terms of Jukebox, I don't think saving legs is a strategy that they really consider. They have a lot of confidence in their top lines, and even when they were down to you, I don't think they thought they were going to lose.

Oh . . . and thanks for saying we don't match up on paper, Kyle. That will be good bulletin board material for next year.

Josh McCarthy said...

At this moment, I would just like to take the opportunity to thank Will Neff, Ryan Purcell, that dude with the incredible game saving block, and the rest of Magnum for fighting to the finish and pulling it out against Texas in pool play 2008. Way to go guys! Thanks for making our tournament.

Josh McCarthy
Harvard Coach 2008

(Also will take responsibility for Boss Hogg tanking against Machine in 2002....but only after the game was starting to get out of hand. Would have still been in our interest to win from the outset - to Tarr's comments - as that would have put us in the winner's bracket with, at the very least, a guaranteed trip to the prequarters. At some point during the course of the game, however, it made sense for us to drop into the bottom pool with a lousy point diff because while we would wind up being technically the 4 seed in the bottom pool, we went in with a 1-0 record, as Machine's margin of victory against us meant that they were elevated to the power pool, and not Epig (who had previously beaten Machine but not by that much). I recall having a discussion w/someone the last year or two that this loophole had since been corrected at Nationals - but based on the comments to this post, it sounds like the answer is no?).

parinella said...

Josh, I think the loophole that was corrected was that the 1-0 team plays the 0-1 team in the first game on Friday instead of #1 vs #2 so that you can't have an 0-2 team play a 2-0 team in the second round in a completely meaningless game.

I have suggested before that the higher seed automatically be given a 1-0 record, but I'm sure there are problems with that, like in tiebreakers and point differentials.

Gambler said...

I've always thought that the way to play to conserve energy is to present the team with specific goals before the game that have nothing to do with winning or losing and then to sub with a wide open rotation.

Those goals could be to work on the team's zone defenses, practice a new clam/junk set, fine tune a different offensive set, iron out your plays, etc. Something that you convince the team they'll need later that tournament or season. Then you are playing with a purpose and don't lose as much focus or mental edge as you would if you go in treating a game as strictly meaningless or if you lose big. You can still tell the team that this is part of a conserving your legs strategy, but getting something else out of the game can really help team psyche--regardless of whether you are conserving in a win or a loss.

Kyle Weisbrod said...

Right on, Gwen. Of course that makes sense and, as a coach, it's how I approach games I know that we are going to win or lose so that the teams stays mentally sharp.

Have you ever effectively employed that in a game where you started with the goal of winning (and subbed accordingly) but at some point in the game realized that winning wasn't worth it?