Spurs late game defense

[ElNono]

somebody tell me they understand my babble

[Mel_13]

Internet just went out in my house. posting this from my phone.

You win.

[ElNono]

[ElNono]

Internet just went out in my house. posting this from my phone.

You win.

Mel, it's not about winning

I just hope you understand what I mean, that's all.

[Mel_13]

Mel, it's not about winning

I just hope you understand what I mean, that's all.

I was joking about winning and I understand your points.
Please don't make me type more from my phone.

[ElNono]

I was joking about winning and I understand your points.
Please don't make me type more from my phone.

[mathbzh]

I was just going to ask if anybody read what was written... teams shooting under 60% at the free throw line? How do you defend that?

I agree with you on that.
But I wonder what your FT% look like if you avoid fouling on good shooters (anybody shooting over 75%) and hack bad ones when possible.

Here is an old link about clutch FT shooting. I guess things have not changed too much.

http://www.82games.com/random23.htm

Thee league average stays around 75%... but considering good shooters tend to shoot more in the clutch there is actually a drop in clutch FT shooting (2.3% according to the link). I don't have the numbers to back it up. But I expect good shooters to stay more consistent (better routine, more experience of clutch FT shooting...).

Now 60% is a reach. But if you try to hack bad shooters and avoid fouling good shooters, you could probably maintain your opponent way under 70%.

[ElNono]

I agree with you on that.
But I wonder what your FT% look like if you avoid fouling on good shooters (anybody shooting over 75%) and hack bad ones when possible.

Here is an old link about clutch FT shooting. I guess things have not changed too much.

http://www.82games.com/random23.htm

Thee league average stays around 75%... but considering good shooters tend to shoot more in the clutch there is actually a drop in clutch FT shooting (2.3% according to the link). I don't have the numbers to back it up. But I expect good shooters to stay more consistent (better routine, more experience of clutch FT shooting...).

Now 60% is a reach. But if you try to hack bad shooters and avoid fouling good shooters, you could probably maintain your opponent way under 70%.

I think this is a situation where there's not really intentional fouling (except for gross cases of 'hack-a-whoever', which the Spurs use more than other teams, but are still extremely rare). As a general rule, the Spurs foul very little. I'm not privy of the foul rate at end of games, but I suspect it's even smaller when they have the lead (which is the presented case). You would have to see if the 60% comes simply from a small sample size (although 5 mins is a lot of time, seems unlikely). The 3 pointers however, we do have the numbers, and I just find it really unlikely you can sustain those percentages for any amount of time outside of just being lucky.

Sometimes you do need a little luck.

[Spursfanfromafar]

Some data and graphs based on play by play info (all calculated using excel and string matching.. there might be errors ) -

All Spurs Games till March 31st -

Opponent ALL FG% in Crunch Time (Below 05:00 remaining in the clock, Turnovers are counted as 2pt FG attempts) -



Opponent 2pt FG% in Crunch Time (Below 05:00 remaining in the clock, Turnovers are counted as 2pt FG attempts) -



Opponent 3pt FG% in Crunch Time (Below 05:00 remaining in the clock) -



Opponent FT% in Crunch Time (Below 05:00 remaining in the clock) -



If required, please put this as a separate post.

[therealtruth]

I think this is a situation where there's not really intentional fouling (except for gross cases of 'hack-a-whoever', which the Spurs use more than other teams, but are still extremely rare). As a general rule, the Spurs foul very little. I'm not privy of the foul rate at end of games, but I suspect it's even smaller when they have the lead (which is the presented case). You would have to see if the 60% comes simply from a small sample size (although 5 mins is a lot of time, seems unlikely). The 3 pointers however, we do have the numbers, and I just find it really unlikely you can sustain those percentages for any amount of time outside of just being lucky.

Sometimes you do need a little luck.

You might not be able to control the percentages but you can definitely control the number of attempts.

[Mel_13]

Some data and graphs based on play by play info (all calculated using excel and string matching.. there might be errors ) -

I really appreciate the effort as this exactly the sort of raw data I was hoping to see. Unfortunately, there are errors. The number of FGAs, even including turnovers, appeared high on their face. I checked the play-by-plays of the two games with most FGAs (Detroit and Denver) and the numbers in the charts are well off from the facts of the game.

Also, when you add the totals from the 2pt FGAs to the numbers from the 3pt FGAs they should equal the numbers for all FGAs. They don't.

I hope you can find the glitch.

Thanks again.

[timvp]

If we had the actual stats, we could figure out how much luck plays into it. I don't see any of these stats in the public domain but I'll send out a few emails . . .

After some begging, got the numbers:

115 minutes of "clutch" defense
224 defensive possessions
77-187 on field goals
9-50 on three-pointers
42-64 on free throws
205 points

I also asked if any of the fouls were of the hack-a-shaq variety and there were six free throws by Ben Wallace. I looked at the play-by-play and Wallace went 3-for-6 in crunch time that game.

Very interesting numbers.

So the Spurs allow opponents to shoot 49.6% (68-137) on two-pointers during clutch situations, which is even worse than their normal allowed two-point percentage of 48.2%. That's not so good, tbh.

Free throw wise, though, if you take out the Ben Wallace attempts, opponents are 39-58 for 67.2%. That's really not too much of a fluke. Add in three more makes and that percentage is up to a normal-ish 72.4%. Looks like I was wrong about the free throw numbers skewing the numbers ... didn't really make much of difference.

But then that three-point percentage of 18% still looks like a fluke, especially since the two-point percentage is so high. A more reasonable number would be 16-for-50 for 32% (the Spurs usually give up 36% on threes but that's factoring the added pressure). So that's seven made three-pointers for a total of 21 points.

Doing the math, 205 points allowed + 3 for the free throws + 21 for the three-pointers = 229 points when adjusted. 229 points allowed in 224 defensive possessions = 102.23 points allowed per 100 possessions. On the season, the Spurs allow 101.4 points per 100 possessions.

Damn, so yeah, unfortunately it doesn't look like the Spurs improve defensively in the clutch. I mean, unless you want to argue that 18% on three-pointers is somehow sustainable, it looks like the Spurs defense remains pretty much the same.




Anyone else have a different opinion regarding those numbers? Hopefully I'm missing something when figuring whether or not this is sustainable . . .

[Spursfanfromafar]

I really appreciate the effort as this exactly the sort of raw data I was hoping to see. Unfortunately, there are errors. The number of FGAs, even including turnovers, appeared high on their face. I checked the play-by-plays of the two games with most FGAs (Detroit and Denver) and the numbers in the charts are well off from the facts of the game.

Also, when you add the totals from the 2pt FGAs to the numbers from the 3pt FGAs they should equal the numbers for all FGAs. They don't.

I hope you can find the glitch.

Thanks again.

Please note that I have added free throws as part of overall FG.

There is still one inconsistency. I forgot to remove the "jumpball" situations from overall plays.

[Mel_13]

After some begging, got the numbers:

115 minutes of "clutch" defense
224 defensive possessions
77-187 on field goals
9-50 on three-pointers
42-64 on free throws
205 points

I also asked if any of the fouls were of the hack-a-shaq variety and there were six free throws by Ben Wallace. I looked at the play-by-play and Wallace went 3-for-6 in crunch time that game.

Very interesting numbers.

So the Spurs allow opponents to shoot 49.6% (68-137) on two-pointers during clutch situations, which is even worse than their normal allowed two-point percentage of 48.2%. That's not so good, tbh.

Free throw wise, though, if you take out the Ben Wallace attempts, opponents are 39-58 for 67.2%. That's really not too much of a fluke. Add in three more makes and that percentage is up to a normal-ish 72.4%. Looks like I was wrong about the free throw numbers skewing the numbers ... didn't really make much of difference.

But then that three-point percentage of 18% still looks like a fluke, especially since the two-point percentage is so high. A more reasonable number would be 16-for-50 for 32% (the Spurs usually give up 36% on threes but that's factoring the added pressure). So that's seven made three-pointers for a total of 21 points.

Doing the math, 205 points allowed + 3 for the free throws + 21 for the three-pointers = 229 points when adjusted. 229 points allowed in 224 defensive possessions = 102.23 points allowed per 100 possessions. On the season, the Spurs allow 101.4 points per 100 possessions.

Damn, so yeah, unfortunately it doesn't look like the Spurs improve defensively in the clutch. I mean, unless you want to argue that 18% on three-pointers is somehow sustainable, it looks like the Spurs defense remains pretty much the same.




Anyone else have a different opinion regarding those numbers? Hopefully I'm missing something when figuring whether or not this is sustainable . . .

timvp with the Paul Harvey.

Thanks.

[100%duncan]

So we didn't really improve? Man the math you're giving me making me head ache.

[Spursfanfromafar]

Have added granularity to Timvp's numbers -

Please note that I have considered Turnovers as Two point attempts. Here is the clutch data (Opposition performance in the last 5 minutes trailing/ leading by less than/equal to 5 points) -



The numbers are off by Timvp's by very few. All data is from play-by-play information.

[Mel_13]

Have added granularity to Timvp's numbers -

Please note that I have considered Turnovers as Two point attempts. Here is the clutch data (Opposition performance in the last 5 minutes trailing/ leading by less than/equal to 5 points) -



The numbers are off by Timvp's by very few. All data is from play-by-play information.

2pt attempts plus 3pt attempts equal total attempts, but 2pt made plus 3pt made do not equal total made.

What am I missing?

[timvp]

So we didn't really improve?

Unfortunately, that appears to be the case. The Spurs are allowing the opposition to shoot 49.6% on two-pointers in clutch situations according to those numbers. That number is too high to claim any sort of improvement, tbh.

[100%duncan]

Unfortunately, that appears to be the case. The Spurs are allowing the opposition to shoot 49.6% on two-pointers in clutch situations according to those numbers. That number is too high to claim any sort of improvement, tbh.

's gotta improve quickly or we will not get to June.

[Mel_13]

Unfortunately, that appears to be the case. The Spurs are allowing the opposition to shoot 49.6% on two-pointers in clutch situations according to those numbers. That number is too high to claim any sort of improvement, tbh.

I'm not completely convinced that you can reliably draw that conclusion. Someone with the time and the skills would really have to analyze the situations and the shot selections.

I just went through a number of the games that meet the criteria. Two things strike me:

1. The trailing team takes a bunch of three pointers.

2. Most of the two point attempts are dunks and layups.

Now this is just an observation from a small sample, but it fits intuitively with what we might expect. Trailing teams will look for an easy two-pointer and then jack up difficult three-pointers if the easy close-in shot is not available.

That could account for a higher than expected 2pt% and a lower than expected 3pt%. (a higher % of 2pt attempts coming from close in than during the rest of the game and a higher % of "bad" or "ill-advised" 3pt attempts than during the rest of the game).

It would be interesting to see a larger, more detailed set of data.

[Spursfanfromafar]

2pt attempts plus 3pt attempts equal total attempts, but 2pt made plus 3pt made do not equal total made.

What am I missing?

2ptA + 3ptA + FTA = Total Attempts

Note that certain teams don't have FTA in the clutch period.

[Mel_13]

2ptA + 3ptA + FTA = Total Attempts

Note that certain teams don't have FTA in the clutch period.

I accept whatever definitions you're using, but when you look at the numbers of "attempts made" the subordinate numbers don't add up to the total.

[timvp]

I'm not completely convinced that you can reliably draw that conclusion. Someone with the time and the skills would really have to analyze the situations and the shot selections.

I just went through a number of the games that meet the criteria. Two things strike me:

1. The trailing team takes a bunch of three pointers.

2. Most of the two point attempts are dunks and layups.

Now this is just an observation from a small sample, but it fits intuitively with what we might expect. Trailing teams will look for an easy two-pointer and then jack up difficult three-pointers if the easy close-in shot is not available.

I would say that is the behavior of teams that are trailing with under a minute remaining. If a team is within five points and there's more than a minute remaining, the offense is going to try to get the best shot possible ... and the defense has to defend both two-pointers and three-pointers. I can't think of a situation outside of a minute remaining and the Spurs up by less than five points where they would just concede a layup or dunk.



Then again, the sample size is so small (115 minutes) that no matter what the numbers say, it's really difficult to draw any conclusions. Even if someone could prove the Spurs have been legitimately great or legitimately lucky, it's very likely a poor indicator of future success/failure.

[Spursfanfromafar]

I accept whatever definitions you're using, but when you look at the numbers of "attempts made" the subordinate numbers don't add up to the total.

I apologise. There was a filtering error and that screwed up calculations for 2FG.

This is the accurate one -



---

Qualitatively speaking, these numbers show very little because there is a lot of noise and the sample size is small - only 21 games where the criteria fits...that is about 105 minutes of play.

However, one can try to make out performances against certain matchups. The Clippers, T'wolves, and the Kings have done the best in clutch situations against the team. It is a case of three point shooting for the Clips, and perhaps frontline depth for the T'wolves and some good interior play for the Kings.

[Mel_13]

I would say that is the behavior of teams that are trailing with under a minute remaining. If a team is within five points and there's more than a minute remaining, the offense is going to try to get the best shot possible ... and the defense has to defend both two-pointers and three-pointers. I can't think of a situation outside of a minute remaining and the Spurs up by less than five points where they would just concede a layup or dunk.

I agree with this, but I wasn't suggesting that the Spurs were conceding layups or dunks (although with this small a sample size a few conceded shots could impact the numbers). I was suggesting the possibility that the offense was taking a much higher percentage of their 2pt attempts right at the rim in these late game situations rather than at other points in the game, and a higher % of their jump shots from beyond the arc than normal. If the ratio of close-in shots to 2pt jump shots, and the ratio of 2pt jump shots to 3pt jump shots, is much different in late game situations, that could help account for the strange combination of stats (high 2pt%, low 3pt%, low pts per possession)

Then again, the sample size is so small (115 minutes) that no matter what the numbers say, it's really difficult to draw any conclusions. Even if someone could prove the Spurs have been legitimately great or legitimately lucky, it's very likely a poor indicator of future success/failure.

Agreed. That's a Tim/Tiago sample size.