Dr. Krzyzewski or: How I Learned to Stop Worrying and Love the Stall

[roywhite]

As the game gets closer, though, the fans get back into it. In Charlotte (UNC country), that made the end very much like a hostile away game and engergized the opponent.

Not to derail this thread, but it's a bit unfortunate to follow the geographical proximity of last year's opponents (Baylor in Houston and Butler in Indy) with more of the same this year, AZ and poss. San Diego St. in Anaheim. That plus the usual underdog factor should lead to a majority hostile crowd this weekend.

Not only is this thread good, the title is superb.

[House G]

It seems to me that there may be some parallels between stall-ball and playing "prevent defense" in football. Although one is an offensive strategy and the other a defensive strategy, they both are used at the end of the game/half to decrease the chances of a successful comeback by the opposing team when the amount of time left is an issue. People question the strategy of stall-ball when it doesn't work. Similarly, it doesn't go over too well when your football team plays lockdown defense all game long only to lose by going to a prevent "D".

[swood1000]

If historically and mathematically stalling gives us the best chance to win, then it makes no sense to deviate from it.

If stalling is appropriate maybe we should add a more aggressive version. Instead of stalling by trying to hold the ball near mid-court we stall by moving the ball around or running plays but restricting the criteria for taking a shot: i.e. only layups or other high-probability shots can be taken until we reach the end of what would be the normal stall period.

Not as safe as a more conservative stall, but at least we're probing the defense and increasing the likelihood of finding something. Also, the team would not be dropping out of offensive mode, which in some cases might cause them to lose their touch.

[Starter]

You can't judge a strategy after the fact. If historically and mathematically stalling gives us the best chance to win, then it makes no sense to deviate from it.

And I assume you read SCMatt33's analysis, but we didn't "[go] away from what was working." Our lead had dropped from 15 points to 8 points before we went into the stall. If we'd played the next five minutes like we played the previous 5 minutes, it would have been a one possession game. Which it was anyway, but it would be much more accurate to say we went away from what was NOT working for something else that didn't work any better.

Or you can look at it as Duke had an eight-point lead and then their next move made the game a toss-up. Seems pretty subjective to assume that a strategy that had for the most part worked most of the half -- even building a 15-point lead at one point -- would completely fall apart in the regular flow of the game if they had continued it. But who knows? Maybe Hardaway continues to evoke terrible visions of Tim Sr.'s Miami Heat days against the Knicks. Anything's possible.

You'll have to take my word for it, but I wouldn't exactly say I judged this strategy after the fact. While having flashbacks of previous tourney games that had slipped away, I was pleading with them to push the tempo as soon as they slowed it down, begging for them to start their offensive sets earlier toward the beginning of SC Matt33's analysis, when they began squandering possessions. (Unlike him, I didn't think any of those possessions were particularly well executed, by the way, though this admittedly is also subjective.)

We can go back and forth all day, but suffice it to say I'm glad we can disagree cordially about a strategic maneuver in a game we won. Lord knows I can only hope we'll be in a similar position after the next four games.

[jv001]

Using the clock(or stall ball) does not bother me, but settling for a 3 with the clock running out actually does bother me. There is more pressure on the shooter in this situation than taking a 3 in the flow of the game. This was particularly true against Michigan with their small lineup. I would have like to have seen Kyle at the high post position in the stall. Not outside taking a 3. I would think we get this worked out this week because Kyrie will probably get more minutes against Ariz. Go Duke!

[Kedsy]

We went away from something that had worked for 29 minutes and didn't work for what... 90 seconds? There's no reason to think it wouldn't have continued to work for the last 8 minutes. We went to something that absolutely didn't "work" for the last 8 minutes. It may have ended up winning the game, but we were grossly outplayed during those last 8 minutes.

As a fan I'd much rather see Duke lose playing their game than lose trying to play the clock. I know that's entirely irrelevant to Coach K and maybe everyone else. There's a chance we lose either way, but one of them would leave a terrible taste in my mouth.

With 10:51 to play we were up by 15. With 8:55 to play we were up by 6. No stalling, we coughed up 9 points of our lead in less than 2 minutes. We went on a mini-rally and built the lead back up to 12 at 7:04, but it was back down to 8 by 5:18 when we arguably went into the stall (although as SCMatt33 points out we took a shot at 4:53 with 10 seconds left on the shot clock). Between 10:51 to play and 5:18 to play we never, not once, got as deep as 20 seconds into the shot clock, so I have no idea what "8 minutes" you are talking about.

I would argue that the time from 10:51 to 5:18 what we were doing was not working. I don't know exactly when Michigan went into the 1-3-1 to stay, but I'd guess it was somewhere around the 10:51 mark (someone please correct me if they know exactly when it happened). So rather than the 90 seconds you suggest, I'd say we had five and a half minutes of not working before the stall. And after five and a half minutes, I think there was plenty of reason to think it might continue to not work for the rest of the game.

[rsvman]

I keep seeing the use of the term "almost lost" in this thread, and in previous threads about stall ball.

The cogent issue here, in my opinion is this: "almost lost"= won.

An "almost loss" is a win. Period. In other words, there's really no such thing as an "almost loss." It's like an "almost miss." If you take a shot from behind the 3-point line, and it bounces around or rolls around the rim and ultimately goes through the basket, what just happened is that you MADE a shot. "Almost miss" is EXACTLY THE SAME as "made." Similarly, "almost lost" is EXACTLY THE SAME as "won."

I would love to have an entire season of "almost losses." That would make us undefeated.

[cato]

It's not like there's a universal consensus that stall ball is the way to go. A vast majority of coaches use it a lot less than Coach K.

I may not agree with your overall argument, but I do think you make some good points -- except this one. If we're going to look at which coaches use which strategies, then we have to look at the relative outcomes. I think K takes all comers on that front.

[InSpades]

One thing that annoys me about how we run the stall... why are we in such a hurry to inbound the ball when we're trying to stall? We go so far as to handicap ourself on offense (be it slightly or greatly) in order to run some more clock and yet we don't take the opportunity to run clock on the inbound. We could easily run 3 seconds before inbounding the ball. And if the other team is not pressuring (like Michigan wasn't) we could run a few more before actually touching the inbounds pass. I'm not sure why no teams use this to their advantage. You could easily get an extra 5-8 seconds per possession without any disadvantage. Am I missing something? This obviously only works after a made basket when the clock is running. I'm assuming the other team would adjust, but this would atleast take them out of what they are trying to do.

[Matches]

I really do think that Kyle having 4 fouls played a role in our going to the stall on Sunday. (And I don't care what K said afterward - I have eyes - that was stallball.) With Kyle defending Michigan's "5", I think he had a greater chance of picking up his 5th foul than he would normally.

I do think there's something to the notion that once a team stops attacking and goes into "prevent" mode, it loses its aggressiveness and opens itself up to a run by the other team. In general, though, I think our team (which is obviously coached to run stallball) is pretty efficient at running it and is equipped to handle the psychological part of it.

Sunday's game, needless to say, was a good example of how NOT to run it. Any strategy executed poorly will look bad in retrospect. In general, though, I think the math bears out that stallball is a reasonable strategy, and much more often than not we seem to run it pretty well.

[NSDukeFan]

We went away from something that had worked for 29 minutes and didn't work for what... 90 seconds? There's no reason to think it wouldn't have continued to work for the last 8 minutes. We went to something that absolutely didn't "work" for the last 8 minutes. It may have ended up winning the game, but we were grossly outplayed during those last 8 minutes.

As a fan I'd much rather see Duke lose playing their game than lose trying to play the clock. I know that's entirely irrelevant to Coach K and maybe everyone else. There's a chance we lose either way, but one of them would leave a terrible taste in my mouth.

But, would you much rather see Duke lose playing their game than win ugly using stalling techniques that mathematically give the team an excellent chance to win? (I realize it's not a fair question.)

If stalling is appropriate maybe we should add a more aggressive version. Instead of stalling by trying to hold the ball near mid-court we stall by moving the ball around or running plays but restricting the criteria for taking a shot: i.e. only layups or other high-probability shots can be taken until we reach the end of what would be the normal stall period.

Not as safe as a more conservative stall, but at least we're probing the defense and increasing the likelihood of finding something. Also, the team would not be dropping out of offensive mode, which in some cases might cause them to lose their touch.

Maybe a more aggressive version would have worked better against Michigan's zone, but I think pfrduke made the point earlier in the thread that if you start attacking earlier, you may also slightly increase the slight chances of a turnover.
http://www.dukebasketballreport.com/...267#post491267

...

We can go back and forth all day, but suffice it to say I'm glad we can disagree cordially about a strategic maneuver in a game we won. Lord knows I can only hope we'll be in a similar position after the next four games.

Wouldn't it be great to have a long discussion about how coach K managed/mismanaged another championship victory in a week and a half? I believe we can all (not to discount kong, shoutingunc, chicagoheel, klemnop, etc.) agree on that.

[Kedsy]

While having flashbacks of previous tourney games that had slipped away...

I hear you but really, how many previous tourney games have slipped away from us after we've gone into the stall? I remember the 1986 championship game and the 1998 Elite Eight game. But Coach K has won 79 NCAA tournament games, and my guess is the vast majority of those wins have featured at least a little bit of stall ball. Once again I go back to the math.

[MChambers]

I would love to have an entire season of "almost losses." That would make us undefeated.

The stress level would be unbelievable, however. We would need advice from heart care specialists.

[InSpades]

I may not agree with your overall argument, but I do think you make some good points -- except this one. If we're going to look at which coaches use which strategies, then we have to look at the relative outcomes. I think K takes all comers on that front.

If you made a list of the most important attributes for a college coach being successful where would you put "end-game strategy with a big lead" on the list? I'm guessing not very highly. Coach is *fantastic* at so many different things. He's one of the best (if not the best) recruiters in the business. He's a fantastic leader. Etc. Just because he's the best coach in college basketball doesn't mean he's the best at every single thing.

[Starter]

I hear you but really, how many previous tourney games have slipped away from us after we've gone into the stall? I remember the 1986 championship game and the 1998 Elite Eight game. But Coach K has won 79 NCAA tournament games, and my guess is the vast majority of those wins have featured at least a little bit of stall ball. Once again I go back to the math.

Well, he's also lost 22 NCAA Tournament games. My guess is the vast majority of those losses...

I'm just playin'. But seriously, get tapes of all this stuff, you and I can have a marathon watching/analysis session. (Let's focus on 2001 and last year, they're more fun to watch.) I have popcorn and surround-sound.

[OldPhiKap]

I keep seeing the use of the term "almost lost" in this thread, and in previous threads about stall ball.

The cogent issue here, in my opinion is this: "almost lost"= won.

An "almost loss" is a win. Period. In other words, there's really no such thing as an "almost loss." It's like an "almost miss." If you take a shot from behind the 3-point line, and it bounces around or rolls around the rim and ultimately goes through the basket, what just happened is that you MADE a shot. "Almost miss" is EXACTLY THE SAME as "made." Similarly, "almost lost" is EXACTLY THE SAME as "won."

I would love to have an entire season of "almost losses." That would make us undefeated.

Yeah, but.

Any strategy/execution which gives the other side the last shot to tie or win is not necessarily a sign of good strategy/execution. We got lucky that the kid missed a seven foot floater, or that they did not kick out for a shot to win it all.

[NSDukeFan]

Well, he's also lost 22 NCAA Tournament games. My guess is the vast majority of those losses...

I'm just playin'. But seriously, get tapes of all this stuff, you and I can have a marathon watching/analysis session. (Let's focus on 2001 and last year, they're more fun to watch.) I have popcorn and surround-sound.

If you focus on 2001 and last year, you won't have a very good opportunity to see how particular strategies didn't work in the NCAA tournament.

[Starter]

If you focus on 2001 and last year, you won't have a very good opportunity to see how particular strategies didn't work in the NCAA tournament.

That's the idea!

[RoyalBlue08]

I know how much the stall ball is hated by many, and I don't think there is any argument that will bring those in that camp around. But I would invite people who watch basketball games other than Duke to keep track of how many games are choked away in the final four minutes when the team that was ahead takes quick shots instead of working the clock down. I see this all the time. And when it is happening, I think...there goes another team losing a game just because they don't understand the situation.

[FellowTraveler]

What would our winning percentage be if we played straight up every time we had a 12 point lead with 5 minutes to go? Would it be more than 91%?

I think there may be other, perhaps better models out there, but to add some context to this discussion, this Ken Pomeroy post suggests a team up 12 with 5 minutes to play has a 96.6 percent chance of winning. That doesn't account for game pace or which team has the ball, and assumes the two teams are evenly matched.

As far as I'm concerned, "Team X wins a lot using Strategy Y" tells us very little. The important question isn't "Does Team X win and employ Strategy Y?" -- it's "Does strategy Y maximize Team X's chances of winning?"

I don't know of advanced studies that have addressed the efficacy of stall-ball in a meaningful way, though I suspect some exist. In the absence of such studies, I believe the following:

1) Teams with significant* late** leads are very likely to win regardless of strategic decisions. Therefore most of the approaches we tend to discuss (begin stalling at 6 minutes vs. 4, or run 20 seconds off the clock rather than 25, etc) are likely to have fairly small marginal benefits or drawbacks (I.e. shifting win probability from 93 percent to 91 percent, or vice versa.)

2) I don't generally think Duke starts to take the air out of the ball too early in the game, but my guess is that the strategy would be (slightly -- see above) more successful if they ran down less time per possession. I'd like to see them initiate the offense at the 15-16 second mark, not 9-12.

3) In my opinion, any attempt to decide when to start holding the ball that is based on figuring out how many net points a team can afford to lose per possession without surrendering the lead should set a minimum end goal of a 4-point lead with 0.0 seconds remaining, rather than a 1-point lead with 0.0 seconds remaining. The reason for this is that "the math," as it has been referred to in this thread, does not account for the possibility of a fluke play/bounce/whatever that leads to a buzzer-beating three by the opposition. If you have a sizable lead, your model for maintaining it should be based around a goal of maintaining a fluke-play-proof lead. Yes, the chances of such a play are small -- but remember, we're talking about small marginal gains to begin with, as a 12 point lead with 5 minutes to play is already quite likely to result in victory.

* Yes, that's vague
** That too