Luck

[SFHoya99]

Here's the funny thing about close games in sports. On average, even the best teams end up pretty close to 50/50 in them. Good teams tend to differentiate themselves by winning more games by when games aren't as close, and it completely makes sense.

Think of teams having a range of performances -- from bad to great. Now take one team whose average performance is better than another -- when they play and both have average performances, the first team wins. When Team A plays great, it always wins, but maybe when Team A plays poorly and Team B plays good, it's a close game. On those days, those teams would tend to split a series pretty evenly. But over a range of performances the better team differentiates itself -- it just isn't in the close games.

This is particularly true in baseball, where end game strategy is minimal and everyone has to bat in order. The only thing that seems to influence whether a team is better at close games in baseball is the one real controllable -- who you have pitching. Since managers can save their best relievers for close games, the teams that tend to be better than 50% in one run games year after year tend to have great bullpens.

In basketball, I think there is much more control. You can shuttle players in and out. On offense, you can run plays and choose who takes the last shots. There's a ton of significant strategy choices to make. One only has to remember our last coach and his troubles with late game strategy to understand that record in tight games in basketball isn't all luck.

With that, here's Coach Thompson's (and his players') record, by margin of victory or loss, over the four years (through Marquette). I left the cupcakes in -- most should fall to the immaterial 10+ so they aren't influencing anything much.

MARGIN	W	L	W%	Cum W	Cum L	Cum W%
1	3	1	0.750	3	1	0.750
2	6	0	1.000	9	1	0.900
3	8	2	0.800	17	3	0.850
4	2	2	0.500	19	5	0.792
5	3	2	0.600	22	7	0.759
6	1	3	0.250	23	10	0.697
7	4	8	0.333	27	18	0.600
8	3	1	0.750	30	19	0.612
9	2	3	0.400	32	22	0.593
10+	65	12	0.844	97	34	0.740
TOTAL	97	34	0.740
OT	5	1	0.833

There's no guarantees out there, but that's a ridiculous record in close games. I find it hard to believe the sample is too small or the data corrupted by us intentionally letting teams back in (there have been too many comebacks for that).

Maybe it is this senior class, or Thompson, or how this set of players respond to Thompson, but this run of "luck" for four years hasn't been luck. Maybe we won't be as good down the stretch in the future -- maybe this was lightning in a bottle -- but I have a hard time pitching 17-3 in one possession games as luck.

[Saxifrage]

Try the numbers without the box chart, which isn't showing up well on either of my browsers.

[JohnnyJones]

Mar 2, 2008 20:43:13 GMT -5 Saxifrage said:

Try the numbers without the box chart, which isn't showing up well on either of my browsers.

I am not seeing anything in terms of the stats, and you teed it up so nicely, I need to see them! Can you repost them somehow?

EDIT: See it now - thanks.

[SFHoya99]

Fixed, I think.

[JohnnyJones]

Mar 2, 2008 20:51:17 GMT -5 SFHoya99 said:

Fixed, I think.

Yes - Thanks. Great stuff (both the numbers and the time it took you to put it together).

[GPHoya]

SFHoya

Many thanks for your thoughtful presentation of statistics on a variety of issues. There is a question that I think you may have addressed in the past about late game fouls with three point leads where a statistical analysis would be very helpful. Of course, with the adjusted foul line the data on how often the defensive team controls a missed free throw may be misleading. Today's UCLA-Arizona ending presented the question and Freeman's position for a Wallace miss highlights the risk of the foul strategy, but my gut says that the odds favor that approach over letting the three-pointer go up. In any event, your posts consistently shed light which is much more valuable than the heat and name-calling that comes from our more emotive contributors. Thanks.

[sleepyjackson21]

Thanks for the info SF. Great stuff. 17-3 in games decided by 3 points or less. That's ridiculous.

[Gold Hoya]

GP-
Ken Pomeroy broke down the statistics earlier this year in the context of the Georgetown vs. West Virginia game.

See www.basketballprospectus.com/article.php?articleid=140 about halfway down the page.

Bruce Pearl actually gave the foul in the Tennessee-Memphis game, which was a great decision given Memphis's poor free throw shooting.

[GPHoya]

Thanks for finding this. I guess there is no actual stat on who rebounds a missed foul shot.

[FLHoya]

Mar 2, 2008 21:53:27 GMT -5 sleepyjackson21 said:

Thanks for the info SF. Great stuff. 17-3 in games decided by 3 points or less. That's ridiculous.

What's more, two of those losses (both three point Ls) came late in the 04-05 season, so it's been a while.

(EDIT: The one point loss was the 06 BET semis. So in the past two seasons we're undefeated in the three point or less games)

Interesting presentation. Anyone else ever find themselves staring at the clock with about 3-4 minutes to go in one of our many close games thinking "How in the world are they gonna pull it out THIS time?" That's happened about 5 times this year already for me.

(And interestingly, I just saw the polar opposite of poise down the stretch as Maryland--at home on Senior Day needing a win for their tournament resume--blew a 20 point lead with 11 mins to go and lost to Clemson by three...and it was not pretty.)

[HoyaSinceBirth]

i disagree FL it was a thing of beauty! LOL maryland chocked so badly.

Dear Maryland:

Where were you last year? We missed you. It isn't the same with out you. Please come home.

Your firend,
The NIT

[LizziebethHoya]

Maryland AND Cuse in the NIT?!? Dreams do come true

[SirSaxa]

Good work SF.

I think it is a matter of:
Poise -- coach sets the tone and the team follows, and we have Wallace
Defense -- we really clamp down in end of game situations and have lowest defensive FG%
Offense -- we run our offense so we are always prepared to get a good shot -- we aren't a run and gun, or drive and dish team, or one on one team, we have a highly sophisticated and proven offense. As a result, the coaching staff doesn't have to draw up something the team doesn't otherwise play. It is the offense we run all game. Maybe with a tweak or two, or adjusted emphasis on a certain player, but not something unfamiliar.

[Saxifrage]

Thanks for the chart. It's very interesting to see these results gathered for comparison.

Sets of numbers always invite cherry-picking of results, and I tempted to say that the fact that the Hoyas are 9-1 over the last four years in 1 & 2 point games isn't obviously meaningful in a statistical sense.

Take another equal-sized range in your chart, games decided by 6 & 7 points: Georgetown is a disturbing 5 out of 16 in those games, or 31%, more than forty points under its overall average of 74%.

Is that as meaningful as the fact that the team is 90% in 1 & 2 point games? Seems like it should be, since it's the same-sized range of possibilities. We could construct some kind of explanation, I guess, but the more likely reason is just the randomness of distribution. The statistician's answer would likely be that the 1 & 2 point games are merely more memorable, not more meaningful.

And yet . . . you're on to the fact that it's getting to be a pretty big sample of games for our seniors, and something feels significant about the team's ability to win close games.

Horse-racing has a term, "on the nod," to describe how one of two essentially even horses wins at the end. On the stride, one goes ahead by a nose, and then on the next stride the other goes ahead. The one that wins--is on the nod--when the finish line comes up ought to be random. But the best jockeys get far more than their random share of the nods. They have some almost magical ability to feel the approaching finish line and get the rhythm of their horses into the winning nod.

The end of a close basketball game is like a horse race won on the nod--the buzzer at the end of the game like the finish line--and this group of seniors and their coach at Georgetown have had too much luck to call it luck at winning close games. That talent doesn't help them when they're six or seven points back, but it pulls out games that, statistically, they should have lost.

[HOYAPLAYA]

Mar 2, 2008 22:07:44 GMT -5 FLHoya said:

Mar 2, 2008 21:53:27 GMT -5 sleepyjackson21 said:

Thanks for the info SF. Great stuff. 17-3 in games decided by 3 points or less. That's ridiculous.

What's more, two of those losses (both three point Ls) came late in the 04-05 season, so it's been a while.

(EDIT: The one point loss was the 06 BET semis. So in the past two seasons we're undefeated in the three point or less games)

Interesting presentation. Anyone else ever find themselves staring at the clock with about 3-4 minutes to go in one of our many close games thinking "How in the world are they gonna pull it out THIS time?" That's happened about 5 times this year already for me.

FL,
Very true. It's a much better feeling to have at the end of a game than the one we used to have where we'd be up with 3-4 minutes to go and wonder "how in the world are we going to blow this damn lead and ruin my entire weekend".

[kchoya]

Pomeroy has a calculation for "luck" (though I can't find an explanation of the methodology). Through the MQ game, he calculates our luck as +0.051, with the National Rank=50, translating to +1.4 wins.

[hoyatables]

Awesome analysis.

It's coaching, pure and simple. Our guys are unbelievably prepared to handle these situations down the stretch, and they don't put themselves in a position to make stupid mistakes; rather, they benefit from the bonehead mistakes of the other teams. This isn't to say that Crean and Wright are bad coaches--far from it--but they do not do as good a job at coaching their players. James should have challenged Wallace, but should not have been that close.

Of those wins and losses, I would love to see a breakdown of how many were come from behind, hanging onto a lead, or just a back-and-forth (as measured from 10 min in the 2nd half). We've won all three ways, so I doubt a discernible pattern would arise, but I do think it would be interesting to see just how many were come from behind wins. That ability -- to never give up -- is a characteristic that few teams consistently possess, and it has to scare the &*&* out of opposing teams and particularly opposing coaches. Duke certainly inspires that fear, and I think we do, too.

[SFHoya99]

Mar 3, 2008 3:43:17 GMT -5 Saxifrage said:

Thanks for the chart. It's very interesting to see these results gathered for comparison.

Sets of numbers always invite cherry-picking of results, and I tempted to say that the fact that the Hoyas are 9-1 over the last four years in 1 & 2 point games isn't obviously meaningful in a statistical sense.

Take another equal-sized range in your chart, games decided by 6 & 7 points: Georgetown is a disturbing 5 out of 16 in those games, or 31%, more than forty points under its overall average of 74%.

Is that as meaningful as the fact that the team is 90% in 1 & 2 point games? Seems like it should be, since it's the same-sized range of possibilities. We could construct some kind of explanation, I guess, but the more likely reason is just the randomness of distribution. The statistician's answer would likely be that the 1 & 2 point games are merely more memorable, not more meaningful.

And yet . . . you're on to the fact that it's getting to be a pretty big sample of games for our seniors, and something feels significant about the team's ability to win close games.

Horse-racing has a term, "on the nod," to describe how one of two essentially even horses wins at the end. On the stride, one goes ahead by a nose, and then on the next stride the other goes ahead. The one that wins--is on the nod--when the finish line comes up ought to be random. But the best jockeys get far more than their random share of the nods. They have some almost magical ability to feel the approaching finish line and get the rhythm of their horses into the winning nod.

The end of a close basketball game is like a horse race won on the nod--the buzzer at the end of the game like the finish line--and this group of seniors and their coach at Georgetown have had too much luck to call it luck at winning close games. That talent doesn't help them when they're six or seven points back, but it pulls out games that, statistically, they should have lost.

Some great points, Saxifrage. I'll try to address them as best I can.

• I don't think you can compare our Winning % in close games to the overall (.740).
  
  For one, that includes a ton of cupcakes, and very few of those are in the "close games" sample. I may still do a Big East + Major Conference + Postseason sample.
  
  Also, just because you have a .700 winning % overall doesn't mean you are likely to have a .700 winning % in close games. In baseball, you're likely to have a 50% winning % in 1-run games regardless of record.
  
  Is that true in basketball? I dunno -- I certainly don't have the data set. I suspect it isn't so clear cut, but I also suspect that even great teams are closer to 50/50 than they are to their overall %.
  
  
• Regarding the dip in the performance later in the range -- namely six and seven point games -- I have no idea where it is just systematic or random distribution. It definitely puts a damper on the 17-3.
  
  That said, I'd submit that even 27-18 in seven point games is pretty damn impressive, especially since most of those are in the BE.
  
  
• Lastly, what I might find most impressive is that easily 2/3 of our games are versus legit opponents, if not more. But we're also winning 2/3 of our wins by 10 or more -- so at least 50% of our legit games in the Thompson era haven't even been close.
  
  Pretty impressive.
  
  

[hoyainspirit]

I remember a post from someone that extolled JTIII's prowess in coaching Princeton to wins in close games. The post was made soon after JTIII was hired. I would say that the poster was absolutely correct.

Thanx for the analysis, SF. Always appreciate your work!

[hoyatables]

Mar 3, 2008 11:22:20 GMT -5 SFHoya99 said:

Regarding the dip in the performance later in the range -- namely six and seven point games -- I have no idea where it is just systematic or random distribution. It definitely puts a damper on the 17-3.

That said, I'd submit that even 27-18 in seven point games is pretty damn impressive, especially since most of those are in the BE.

I think that when we do lose games, it tends to be more decisive. I suspect that many of those games were ones where we were just out of it from the start, or had a big lead build up on us in the second half, or something along those lines. It's not that we failed to close the gap on a 6-8 point deficit in the closing minutes; rather, the game was well out of our hands by the 10 min mark in the second half and we were only able to close it to within 6-8 points by the buzzer.

I would love to take the time to break the wins and losses down even further and evaluate how productive we are when down at the half, or at the 10 min mark, or at the 5 min mark. That's a good plan for the offseason I think. This is a very interesting discussion that is well worth continuing !