Why Lickliter Plays So Slow (A Theory)

[Bumped, and quite worth a read.--OPS]

In a recent post, StoopsMyAss described the "Lickliter way".  This post just adds one facet to the picture he painted (I commented on SMA's post, and this is just an extension of that post).  The most striking aspect of the "Lickliter way" for me has always been the pace of the game.  His teams play slooooow.  This is subjectively obvious to anyone who watches a game, but it's objectively true, too.  Here's how his teams have ranked in terms of possessions per game for the past six years:

08-09: 56.7 poss/game (344th out of 344 teams)
07-08: 59.5 poss/game (334th out of 341 teams)
06-07 [Butler]: 58.4 poss/game (332nd out of 334 teams)
05-06 [Butler]: 59.4 poss/game (327th out of 332 teams)
04-05 [Butler]: 58.1 poss/game (327th out of 330 teams)
03-04 [Butler]: 57.9 poss/game (325th out of 326 teams)

That is really slow!  Lickliter's teams are regularly in the bottom five in the nation in terms of pace.  For a sense of perspective, the average pace is 66.5 poss/game.  As I watched Iowa play, I began to wonder: what does Lickliter hope to accomplish by this?  And I wasn't (just) asking out of frustration.  I assumed there must have been some good reason to play this way.  After all, other teams gained some measure of success playing at a slow pace, Butler being a prime example.  The usual reasons broadcasters and sportswriters gave -- that the slow pace "frustrated" other teams into making mistakes, that we would use our possessions more efficiently -- weren't very satisfying.  "Are teams really that frustrated by playing slowly?", I thought.  "And can't we only control the pace of our own possessions, not the other team's?  Shouldn't it be our team, not theirs, that gets frustrated by playing slow?  And if our offense is so dang efficient, why not give it more opportunities in a game to do its efficient-tastic thing?"  With these questions in mind, I went to Google, and found an answer.  It's not necessarily the answer, but it does explain one way that playing slow helps a team compete.  And this answer comes down to what is, depending on your view, either the savior or the anti-Christ of modern sports: statistics.


What I found was an article by Dean Oliver* entitled "The Effect of Controlling Tempo".  I recommend reading it if you have the time (it's not long, although it is a bit math-heavy).  I'm not adding anything to his reasoning, just summing it up, but I will talk a little bit about what the implications of Oliver's argument are for Iowa.  Oliver's argument is basically this: slowing down the pace of the game increases the variability of the game's outcome, which improves your chances of winning if you're the underdog.  Oliver gives a detailed mathematical explanation for why this turns out to be so, but you can understand it intuitively by looking at an example from outside of sports.  Imagine the following (made-up) card game: each turn, you draw a card from a deck.  If the card is a face card, you get a point.  If it's not, your opponent gets a point.  At the end of a series of turns (with you drawing the card every turn), you count up the points, and whoever has the most points gets $100 from the other guy.  This may sound like a pretty crappy game for you.  A little math shows that your chances of drawing a face card are less than one in four, which puts you at a decided disadvantage from the get-go.  Most people would prefer to decline the game altogether, but suppose that you were forced to play, possibly by some mathematically-oriented Jigsaw type, and that you were also forced to make a choice: either play a short game (lasting 11 turns) or a long game (lasting 101 turns).  Which would you choose? 

* No, not
that Dean Oliver.  This Dean Oliver is a big name in the world of basketball statistics, the author of a great book on the subject, "Basketball on Paper", and currently an adviser to the Denver Nuggets.

If the short game seems more appealing, there's a reason.  It's much more probable that you'll get lucky and win six (or more) out of eleven turns than that you'll win 51 (or more) out of 101.  If you could play this game for an infinite number of turns, you would certainly lose the overall game, because you would end up winning very close to 3/13 of the turns and losing very close to 10/13 of the turns.*  If you play a finite number of terms, however, the percentage of turns you win will vary a little.  That is, you will win a few times more or less than expected.  The fewer the number of turns you play, the more variability you will see in your turn-winning average.  If you reduced the game to one turn, you would either average one win per turn or zero wins per turn, both of which are pretty far from what we know your expected average to be: 3/13.  And in this game, it's good for you to vary from the expected average, because the expected average means you lose.  In a game that lasts just one turn, you have a 3/13 (23%) chance of winning.  If you extend the game to just 11 turns, your chances of winning drop to 8.2%.  If you extend it to 101 turns, your chances of winning go down to .000000000081% (about one in a trillion).

* To be exact, your winning percentage would approach 3/13 as the number of turns approached infinity.  This is due to something called the law of large numbers.

What does this have to do with basketball?  Well, basketball is kind of like the made-up game I just described.  In a basketball game, two teams take turns (i.e., possessions) trying to score, and we add up the total number of points at the end to determine a winner.  Just as the above game favors one player, basketball games usually favor one team.  That is, one team is "better".  What do I mean by "better"?  I mean that if the two teams could somehow play an infinite number of possessions without getting tired, one team would score more points per possession than the other.*  In other words, one team has better odds of scoring each time it has the ball.  Since each team gets roughly the same number of possessions**, the team with the better odds of scoring each possession should win the game.  As we've seen with the card game, however, it's not that simple.  If you slow the game down really slow and thus reduce the number of "turns" each team gets, weird stuff starts to happen.  A turn of events that might not matter very much in a 100-90 game, like a missed three by one team followed by a made three by the other, can make a huge difference in 50-45 game.  There are so few possessions that a small run of good possessions by one team and bad possessions by the other can decide the game.  And for the worse team, that's a good thing.  The worse team wants weird, it wants lucky, because it know that the longer the game goes on, the more luck will even out for both sides. 

*This may seem like an unusual definition of "better", since most sports fans would say that the "better" team is the one that wins a particular match-up, but it's the definition that a statistician would favor.  The problem with taking one particular observation (i.e., one game) as "proof" of the superiority of one team over another is that anything can happen once.  Two aces make for a "better" starting hand than a two and a seven off-suit in No-limit Hold'Em, but that doesn't mean that two aces will
always win.  In 1985, Georgetown was probably "better" than Villanova, but we all know how that ended.  The nature of sports makes it hard to repeat games as often as we would like to say, "Team A is definitely better than Team B", but people like Oliver have figured out ways to get around this handicap.  Namely, they have observed that the best way to predict if one team will beat another is to pay attention to two little numbers: points scored per possession and points allowed per possession.  Good teams score more points per possession than they give up, mediocre teams break even, and bad teams get outscored.  As long as two teams play roughly the same competition, points scored per possession minus points allowed per possession is a good indicator of relative team strength.  The "per possession" part is important, because it allows us to compare apples (teams that use lots of possession per game, i.e. play fast) to oranges (teams that use few possessions per game, i.e. play slow). 
**Every time one team loses possession, whether by making a shot, missing a shot or turning the ball over, the other team gains a possession, so possessions usually match up pretty closely.  One reason they don't match up perfectly is that one team ends the half with the ball, thus giving them an "extra" possession.

This tactic doesn't guarantee victory.  You have to play well and you have to be lucky for it to pay off.  Some teams have used an extreme form of this strategy in the NCAA tournament to pull off some major upsets.  For example, Princeton beat #4 seed UCLA 43-41 as a #13 seed in 1996, and almost beat #1 seed Georgetown in 1989 (they lost 50-49).  As great a victory as that 1996 game was for Princeton, you'll note that they didn't go much further in the tournament (they lost in the next round to Mississippi St.).  Your shots need to be falling, and the other team needs to go cold.  All this tactic does is help to magnify the importance of small runs of luck.

It's also important to note that favored teams can achieve the opposite effect by playing fast.  If a team is aware of the way a slow pace helps the underdog, they can counteract this effect by speeding the game up on their end.*  It's kind of like if Peja Stojakovic had a three-point shootout against Shaquille O'Neal and had to decide whether to make the contest last one shot or one hundred shots.  Since Peja knows he's the better shooter, he wants to have more chances to prove this superiority.  What he doesn't want is to lose to Shaq based on one lucky make and miss.  You can see this tactic employed by the teams that know they "should" win, that is, teams with massive advantages in skill and athleticism.  A good example came in Iowa's game with Texas the other day.  I don't have a breakdown of possessions by half, but it seemed to me that Texas made a conscious effort to push the tempo in the second half.  I would argue they did so at least partially to avoid the dangerous runs of luck a slow game features.  North Carolina often utilizes the same strategy.  They know they have the better team and they try to give their team as many chances to demonstrate that as possible (or maybe they just like to run fast and dunk a lot -- they could win any way they like, I imagine).

*There's a limit to how much you can usefully speed up the game, of course.  The Phoenix Suns famously tried to score in "seven seconds or less".  A "two seconds or less" strategy would have probably resulted in a lot of low-percentage half-court heaves.

The Texas game is also a good example of how the slow strategy can work for Iowa.  For much of the first half things weren't going Iowa's way, but a string of made threes (capped by a spectacularly lucky full court heave by Cully Payne) brought them back to tie the game.  If the game had ended there, we would have been in great shape.  Unfortunately, we had to play 20 more minutes, and our luck evened out. 

You may ask here, "this could all be true, but does any coach actually think about it?"  Maybe so, maybe not.  It could be that a coach comes to the slow approach through a trial and error approach.  He might just like playing slow as a personality quirk.  Then he might notice that this approach seems to work for the talent on his team and decide to slow down even more.  Then he might have a eureka moment and say, "if slow is good, and slower is better, then slowest must be best!"  And voila, that coach would play very slow.  Who knows.  It's hard to read too much into a coach's psychology, so I think it's safer to focus on what we can observe, and that's that Iowa does play very slow and that this slow pace does have an effect on our results.

What is this effect?  Essentially, playing slow gives us better chances when we're the underdog and worse chances when we're not.  Since we've frequently been the underdog the past few years, this strategy has probably helped us (it may, in fact, be one reason Lickliter has been reluctant to push tempo).  Maybe when our talent level improves, Lickliter will speed things up.  Don't hold out too much hope, though.  As I mentioned above, Butler was always a very slow team, even when they went to the Sweet 16.  For better or worse, this is Lickliter's style, and he'll probably stick with it.  The variability of the result is only one part of the style, of course.  The other parts, which SMA mentioned in his article -- reducing turnovers, making threes, making free throws, playing good defense -- are much more important, and much more repeatable from game to game.  The Lickliter way is to put an enormous amount of pressure on a small number of possessions, and hope that our team, due to its superior discipline and training, will be better able to utilize those possessions.  It's not Showtime, but it can work.  The fun of this season will be to see whether the players can master the Zen discipline it takes to succeed in this unorthodox system.

Brief editorial post-script

I hope I've shown how playing slow can help an underdog win.  It's another question whether this is an aesthetically pleasing style of basketball.  The main reason we have a shot clock in basketball is that slowing the game down had become such a popular tactic that it was ruining the game (a 1950 game between the Pistons and the Lakers ended 19-18, for instance).  The NCAA instituted a 45-second shot clock in 1985 and reduced it to 35 seconds in 1993, and aesthetics surely had something to do with those decisions.  Beyond a certain point, most people agree that slow basketball is boring basketball, and my opinion is that Iowa has passed that point.  Based on attendance records, most fans seem to agree.  As a fan, I wish we would play a little faster.  I'm not asking for anything crazy, but I could definitely do without deliberate time-wasting maneuvers, such as passing the ball around the three-point line for 30 seconds.  Besides, if what I said about playing slow is true -- that it magnifies small runs of luck -- is that how we want to win (or lose)?

Unless otherwise expressly indicated by BHGP editors, this FanPost is strictly the viewpoint of the author and is not endorsed by BHGP in any way.