PG Assist Breakdown - 3 PT/2 PT/Close

[spectre]

Found this on the Raps' board on RGM.

There's assists and then there are assists. Tossing the ball to a go-to guy where he shoots a contested shot on the perimeter is much different then actually creating the shot. Though this doesn't specifically do that (and even with jump shots the PG could easily have driven, pulled the other defender to him and "created" space for the jump shooter), it does give an interesting perspective.


Player Name 3P 2Jump Close Total %3P %2Jump %Close
Mo Williams 62 73 231 366 16.94% 19.95% 63.11%
Jamaal Tinsley 14 31 63 108 12.96% 28.70% 58.33%
Raymond Felton 116 90 240 446 26.01% 20.18% 53.81%
Derek Fisher 33 60 108 201 16.42% 29.85% 53.73%
Devin Harris 75 121 227 423 17.73% 28.61% 53.66%
Anthony Carter 36 40 84 160 22.50% 25.00% 52.50%
Mike Bibby 43 107 159 309 13.92% 34.63% 51.46%
Baron Davis 108 187 303 598 18.06% 31.27% 50.67%
Marcus Williams 12 69 80 161 7.45% 42.86% 49.69%
Lou Williams 59 78 131 268 22.01% 29.10% 48.88%
Mike Conley 75 145 205 425 17.65% 34.12% 48.24%
Deron Williams 156 259 384 799 19.52% 32.42% 48.06%
Steve Nash 218 250 424 892 24.44% 28.03% 47.53%
Jose Calderon 91 125 188 404 22.52% 30.94% 46.53%
Rajon Rondo 186 240 368 794 23.43% 30.23% 46.35%
Aaron Brooks 117 116 201 434 26.96% 26.73% 46.31%
Stephen Curry 110 144 218 472 23.31% 30.51% 46.19%
Ty Lawson 47 63 94 204 23.04% 30.88% 46.08%
Kyle Lowry 103 62 140 305 33.77% 20.33% 45.90%
Brandon Jennings 137 120 213 470 29.15% 25.53% 45.32%
Chauncey Billups 94 131 183 408 23.04% 32.11% 44.85%
Jameer Nelson 144 53 156 353 40.79% 15.01% 44.19%
Jrue Holiday 53 104 123 280 18.93% 37.14% 43.93%
Gilbert Arenas 38 91 100 229 16.59% 39.74% 43.67%
Jarrett Jack 85 149 179 413 20.58% 36.08% 43.34%
Ramon Sessions 66 82 111 259 25.48% 31.66% 42.86%
Russell Westbrook 138 236 278 652 21.17% 36.20% 42.64%
Darren Collison 107 141 184 432 24.77% 32.64% 42.59%
D.J. Augustin 41 70 82 193 21.24% 36.27% 42.49%
Steve Blake 101 118 161 380 26.58% 31.05% 42.37%
Beno Udrih 80 137 155 372 21.51% 36.83% 41.67%
Eric Maynor 62 94 110 266 23.31% 35.34% 41.35%
Jordan Farmar 36 39 52 127 28.35% 30.71% 40.94%
Andre Miller 114 149 182 445 25.62% 33.48% 40.90%
T.J. Ford 61 46 73 180 33.89% 25.56% 40.56%
Rodney Stuckey 71 140 136 347 20.46% 40.35% 39.19%
Chris Duhon 122 108 148 378 32.28% 28.57% 39.15%
C.J. Watson 63 47 70 180 35.00% 26.11% 38.89%
Jonny Flynn 96 123 138 357 26.89% 34.45% 38.66%
Earl Watson 151 94 154 399 37.84% 23.56% 38.60%
Rafer Alston 31 78 67 176 17.61% 44.32% 38.07%
Chris Paul 132 167 181 480 27.50% 34.79% 37.71%
Nate Robinson 48 55 60 163 29.45% 33.74% 36.81%
Will Bynum 74 107 103 284 26.06% 37.68% 36.27%
Randy Foye 55 92 81 228 24.12% 40.35% 35.53%
Jose Juan Barea 64 103 92 259 24.71% 39.77% 35.52%
Earl Boykins 48 66 57 171 28.07% 38.60% 33.33%
Derrick Rose 105 208 156 469 22.39% 44.35% 33.26%
A.J. Price 45 26 35 106 42.45% 24.53% 33.02%
Luke Ridnour 97 125 103 325 29.85% 38.46% 31.69%
Tyreke Evans 131 151 130 412 31.80% 36.65% 31.55%
Jeff Teague 27 57 38 122 22.13% 46.72% 31.15%
Carlos Arroyo 60 98 65 223 26.91% 43.95% 29.15%
George Hill 65 98 63 226 28.76% 43.36% 27.88%
Toney Douglas 52 30 30 112 46.43% 26.79% 26.79%
Jerryd Bayless 47 84 41 172 27.33% 48.84% 23.84%

This is for the previous season.

[SCBobcat]

So...this is telling us that Felton was best at setting up a man for a bucket in the paint. Also it tells us that Felton sucked at setting up a man for a midrange jump shot. DJ, on the other hand, set his players up for midrange jumpers at a much higher rate, while still feeding the ball to scorers in the paint for over 42% of his assists. Interesting. But what would be our ideal % mix?

[SWedd523]

So...this is telling us that Felton was best at setting up a man for a bucket in the paint. Also it tells us that Felton sucked at setting up a man for a midrange jump shot. DJ, on the other hand, set his players up for midrange jumpers at a much higher rate, while still feeding the ball to scorers in the paint for over 42% of his assists. Interesting. But what would be our ideal % mix?

For our team, since we don't have any jump shooters, I'd look for the close range % to be high.

[teej]

For our team, since we don't have any jump shooters, I'd look for the close range % to be high.

This.

Jack and DJ are the rotation players who can consistently knock down threes, and Jack does that better on his own rather than via assist.

[SCBobcat]

Realistically (and to simplify the math), we as a team might shoot 35% from 3; 60% from the paint; 40% on midrange jumpers; and 75% from the free throw line. If LB will have no more than 20 3 point attempts per game and encourages us to drive the lane, I imagine we would shoot about 35 times from the paint and about 15 midrange jumpers. Without bogging you down with the math, at the above percentages, we would score our points in this fashion: 20.5% from 3; 61.7% in the paint; and 17.6% from midrange. Factor in around 30 FT attempts and we score 96 points per game on 70 attempts from the field (48% shooting from the field overall). Obviously not all baskets are assisted, but it is plain to see that with our type of team philosphy it helps us to have a PG who primarily creates in the paint

[teej]

20.5% from 3

Egad.

Can Matt be our 6th man?

[SCBobcat]

Egad.

Can Matt be our 6th man?

That's not a shooting percentage, that is merely a guess at what percentage of our shots made per game come from behind the arc. A clearer way to say it is if we make 34 shots in a game, I guess that 7 or them (or 20.5%) are 3 pointers taking into account an estimated 20 3s taken at a 3pt shooting percentage of 35%.

[teej]

That's not a shooting percentage, that is merely a guess at what percentage of our shots made per game come from behind the arc. A clearer way to say it is if we make 34 shots in a game, I guess that 7 or them (or 20.5%) are 3 pointers taking into account an estimated 20 3s taken at a 3pt shooting percentage of 35%.

I know. But we need more than 1 out of every 5 shots assisted to be a 3. A lower number of attempts to begin with because of Larry, and no great shooter, means that it's probably even less than that.

The Magic were closer to 2 of 5.

Even with Ray it was 1 of 4.

The 3 ball is important to have success, especially when you don't have an elite wing.

[mrfargo]

Sorry guys, I am learning how this works...lol

[mrfargo]

Realistically (and to simplify the math), we as a team might shoot 35% from 3; 60% from the paint; 40% on midrange jumpers; and 75% from the free throw line. If LB will have no more than 20 3 point attempts per game and encourages us to drive the lane, I imagine we would shoot about 35 times from the paint and about 15 midrange jumpers. Without bogging you down with the math, at the above percentages, we would score our points in this fashion: 20.5% from 3; 61.7% in the paint; and 17.6% from midrange. Factor in around 30 FT attempts and we score 96 points per game on 70 attempts from the field (48% shooting from the field overall). Obviously not all baskets are assisted, but it is plain to see that with our type of team philosphy it helps us to have a PG who primarily creates in the paint


I am so lost!!!!!

[Chef]

I am so lost!!!!!

these stats make sense based on the make up of our team. basically, it shows where our assists come from. it shows that we have no shooters which is what we need to space the floor. jax is a volume shooter (kinda like antoine walker but more accurate) but not a knock down shooter/spacer. which is why getting someone like ben gordon could be beneficial for us even though he isn't the best 3pt shooter and is overpaid.

from the looks of all the point guards, i think this chart basically describes the team first and then whether the pg is one who breaks down defenses off the dribble or more like a distributer like derrick fisher or mario chalmers (strong sg who controls the ball more).

[spectre]

these stats make sense based on the make up of our team.

As these stats are for a lot/most of the players listed.

This stuff is fun to look at and you might be able to get something from them...but like almost everything else they're not stand alone. Who they're playing with and the teams' offensive concept absolutely play a role in how these come out.

To me one of the biggest "flags" on this would be if the assists to 3 PT shots were the highest...but even then we wouldn't know if that was due to just a ball dominant passing it to a 3 pt specialist or if the player actually created the opportunity.

[teej]

To me one of the biggest "flags" on this would be if the assists to 3 PT shots were the highest...but even then we wouldn't know if that was due to just a ball dominant passing it to a 3 pt specialist or if the player actually created the opportunity.

You'd have to match up the player's entire career along with the histories of his teammates.

[spectre]

Wish I could sort these other than by "close".

Something that does jump out at me? The lowest (going to highest) for the 2 PT shots:

Jameer Nelson - 15.01%
Mo Williams - 19.95%
Raymond Felton - 20.18%
Kyle Lowrey - 20.33%
Earl Watson - 23.56%
Anthony Carter - 25.00%
TJ Ford - 25.56%

Does/should that be telling us anything? I can guess on Nelson as his team's strategy is to shoot the 3 (and he has 40.79% in that category). What about the other guys?

[dvdbumpus]

Don't really get this, but anything with Jamaal Tinsley and Mo Williams at the top of PG rankings make this null and void.

[Ampsportsduo]

Don't really get this, but anything with Jamaal Tinsley and Mo Williams at the top of PG rankings make this null and void.

I know it's tempting, but don't do this. Statistical analysis isn't alchemy, you have to keep going back to the drawing board (or excel) over and over to get a valuable product. The focus should be on identifying flaws of the product and looking to improve it. I know it's a joke, but these discussions could be great fodder and help all of us increase our basketball knowledge. I'd hate to see that lost.

While this list is interesting, it doesn't tell the whole story so it's like half of the rosetta stone (or maybe just half of an idea written on a bar nap that should've been thrown away).

For instance, Mo Williams high percentage of "close baskets" can be very deceiving. A dump down to Shaq can equal an assist. A pass from the top of the key to LeBron cutting to the hoop, then a crab dribble and dunk can equal a "close assist." I suspect that this was gleaned from game logs using purely the description of the made basket to determine where the assist occurred. In all likelihood for most cases it says more about the PG's supporting cast than it does the PG (I believe someone said this).

What would be more valuable (imo) is a chart that showed where the PG was on the floor when he made each assist. It would look like a "hotspot" chart if you will and give a better idea of how a guy is breaking down the defense. This would likely have use for both player evals and scouting during the season.

I don't know about a lot of folks on here, but I eat this stuff up. I'd love to see a new 'statistical analysis' section on BCP with a focus on things like this. H/T Spectre for starting this discussion.

[Chef]

I know it's tempting, but don't do this. Statistical analysis isn't alchemy, you have to keep going back to the drawing board (or excel) over and over to get a valuable product. The focus should be on identifying flaws of the product and looking to improve it. I know it's a joke, but these discussions could be great fodder and help all of us increase our basketball knowledge. I'd hate to see that lost.

While this list is interesting, it doesn't tell the whole story so it's like half of the rosetta stone (or maybe just half of an idea written on a bar nap that should've been thrown away).

For instance, Mo Williams high percentage of "close baskets" can be very deceiving. A dump down to Shaq can equal an assist. A pass from the top of the key to LeBron cutting to the hoop, then a crab dribble and dunk can equal a "close assist." I suspect that this was gleaned from game logs using purely the description of the made basket to determine where the assist occurred. In all likelihood for most cases it says more about the PG's supporting cast than it does the PG (I believe someone said this).

What would be more valuable (imo) is a chart that showed where the PG was on the floor when he made each assist. It would look like a "hotspot" chart if you will and give a better idea of how a guy is breaking down the defense. This would likely have use for both player evals and scouting during the season.

I don't know about a lot of folks on here, but I eat this stuff up. I'd love to see a new 'statistical analysis' section on BCP with a focus on things like this. H/T Spectre for starting this discussion.

no matter what the statistical model, a point guard is only as good as his cast. with the exception of the top 5-7 in the league. they make the team. cp3, williams, nash etc. they put up numbers and make players better by themselves. players like mo williams and mario chalmers live off of a superstar. i think that ray was surprisingly good stat wise given his cast. not much movement offensively, no post game and a general lack of knock down shooters. very tough to put up numbers unless you are pretty good floor general.

[spectre]

I know it's tempting, but don't do this. Statistical analysis isn't alchemy, you have to keep going back to the drawing board (or excel) over and over to get a valuable product. The focus should be on identifying flaws of the product and looking to improve it. I know it's a joke, but these discussions could be great fodder and help all of us increase our basketball knowledge. I'd hate to see that lost.

While this list is interesting, it doesn't tell the whole story so it's like half of the rosetta stone (or maybe just half of an idea written on a bar nap that should've been thrown away).

For instance, Mo Williams high percentage of "close baskets" can be very deceiving. A dump down to Shaq can equal an assist. A pass from the top of the key to LeBron cutting to the hoop, then a crab dribble and dunk can equal a "close assist." I suspect that this was gleaned from game logs using purely the description of the made basket to determine where the assist occurred. In all likelihood for most cases it says more about the PG's supporting cast than it does the PG (I believe someone said this).

What would be more valuable (imo) is a chart that showed where the PG was on the floor when he made each assist. It would look like a "hotspot" chart if you will and give a better idea of how a guy is breaking down the defense. This would likely have use for both player evals and scouting during the season.

I don't know about a lot of folks on here, but I eat this stuff up. I'd love to see a new 'statistical analysis' section on BCP with a focus on things like this. H/T Spectre for starting this discussion.

I like this stuff too, and I always look forward to your comments whenever I can find it to post.

I think a separate section would be a great idea! Most times it can at least be interesting; the main problem I've always had with "stat gurus" is they sometimes think/act like these analyses are the "end all/be all" and it trumps what actually happens at the games.

Just like our great posters have noted in this thread...with stats there's almost always more influencing the numbers than what the author of said stats is trying to imply that they mean.

In the end you HAVE to watch the game. Combine that with the stats/breakdowns and then you can work towards getting the total picture.

[spectre]

BTW...as I said I pulled this off the Raps' board on RGM. Their fans used this to come to the conclusion that Jose is really in the "upper reaches" in regards to feeding the paint.

I commend the guy for taking the time to compile this as I'm sure it took a lot of legwork (like Amp said the guy probably went thru the game logs)...but they are just another example of fans trying to convince themselves of a "truth" while ignoring all the contributing factors.

[BlockParty]

I don't know about a lot of folks on here, but I eat this stuff up. I'd love to see a new 'statistical analysis' section on BCP with a focus on things like this. H/T Spectre for starting this discussion.

Agreed! Stats help remove the Sports Center element of our opinions. They justify why you give Okafor a big contract (one of 5 multi-year double double players in the league).

Another idea for PG effectiveness could be matching the PG's stats versus the opposing team's PG's stats....and taking the difference (netting the two together). That combines the PG's defensive capabilities (in most cases, teams don't always defend the same player that is defending you). Afterall, what better measure for a PG than the net impact they have on a game. What this would probably provide is who the top PG's are in the league, it interests me because I don't know what the results would end up being. This combined with minutes actually played together at the same time would add an additional complication.

[Chef]

I like this stuff too, and I always look forward to your comments whenever I can find it to post.

I think a separate section would be a great idea! Most times it can at least be interesting; the main problem I've always had with "stat gurus" is they sometimes think/act like these analyses are the "end all/be all" and it trumps what actually happens at the games.

Just like our great posters have noted in this thread...with stats there's almost always more influencing the numbers than what the author of said stats is trying to imply that they mean.

In the end you HAVE to watch the game. Combine that with the stats/breakdowns and then you can work towards getting the total picture.

send an invite to bret haline.

[Chef]

Agreed! Stats help remove the Sports Center element of our opinions. They justify why you give Okafor a big contract (one of 5 multi-year double double players in the league).

Another idea for PG effectiveness could be matching the PG's stats versus the opposing team's PG's stats....and taking the difference (netting the two together). That combines the PG's defensive capabilities (in most cases, teams don't always defend the same player that is defending you). Afterall, what better measure for a PG than the net impact they have on a game. What this would probably provide is who the top PG's are in the league, it interests me because I don't know what the results would end up being. This combined with minutes actually played together at the same time would add an additional complication.

if you are looking at offense only another way would be to do what they do for one of the baseball stats where basically they compare a player to the average replacement player. in other words, take the point guard's supporting cast and define how good they are so that each point guard's stats can be viewed as though they have the same cast. how you do this, no clue. this way if you have 2002 shaq on your team we can compare your assist numbers to a point guard with shawn bradley as far as "inside assists" go. you could also identify each teams strengths ie jameer nelson gets x number of assists off of drive and kick 3 pointers but the magic lead the league in 3 pt fg's per game and 3 pt fg %. again, the attempt is to qualify a pg's stats based on team strengths. not at the detriment of the pg, just to show the makeup of the team.

[spectre]

send an invite to bret haline.

He used to come on here occasionally. I always check his site.