Tuesday, October 30, 2012

Who’s Shooting The Ball, And From Where?

I had a bit of free time this offseason, so I spent a portion of it combing through box score play-by-play data (I can tell you’re interested already) to see if I could compartmentalize shot selection. The traditional box score has “buckets” for three pointers and total field goals, so it’s easy to separate out into two-point shots and three-point shots, but I wanted to try to unearth another bucket or two, because an 18 foot two-point jump shot is different than a layup in terms of makeability (spell check is telling me this isn’t a word, but I just typed it, and you understood it, so it must be).

I noticed in Calvin’s play-by-play data that there are four types of shot labels: three-pointer, jumper, layup, and dunk. So, I thought the easiest thing to do was to load each game’s play-by-play section into Excel and have it count made and missed “jumpers” for each player for me. This effectively separated mid-range shots from close to the basket attempts and, along with the readily available three-point numbers, gave me three total shot-type buckets to work with. I would have liked to separate out jump shot range further – perhaps something as simple as in the paint and out of the paint – but the data only labeled as shot as in the paint ([PNT]) if it was a made basket, so that idea didn’t work.

Of course, the data is not perfect (technically I’m supposed to write the data are not perfect, but I find that to result in an unnecessarily snooty tone) because it’s entered manually by humans at the scorer’s table who are trying to watch and do about seven hundred things at once, so some of the labels aren’t perfect. For example:
00:04 63-64 V 1 GOOD! LAYUP by Ryan Krombeen [PNT]

Sorry to dig up the bad memory (sad face), but it served a purpose.

I’m not sure I would have necessarily called that a layup myself, but that’s how it was recorded, and we’ll live with it. It’s not really the type of mid-range jumper I was looking to separate out either, so it’s probably not a big issue. I can imagine that a type of observational bias may exists here whereby made shots of this type are more likely to be dubbed “layups” and missed shots are more likely to be “jumpers” (because a made shot looks easier). That’s not a knock against the folks that enter the data, these plays have to be made with real-time split second decisions, but it should be noted that the play-by-play data isn’t gospel. (One often sees this phrase with gospel meaning something like “divine truth”, as I just used it, but the word literally means something like “good news”, so these statements never really make any sense.)

Anyway, I should probably present some numbers before this post reaches the 500 word mark and you, the reader, go all close window, delete bookmark on me. Here are the combined team numbers for the 2012 season:




Shot Type
FGA
eFG%
All Shots
1592
0.474
Close
587
0.538
Mid-range
685
0.406
Three pointer
320
0.502

I use effective field goal percentage instead of standard field goal percentage to put three-point shots on equal footing with non-three point shots (you get 50% more points for making one!).

Obvious conclusion: mid-range jump shots suck. They’re high-risk, low reward shots, and really should only exist as a very minor part of the game plan. They’re necessary on occasion as end of shot clock options and as a way to stretch the defense and keep them honest, but it’s not an efficient way to score points. A three-pointer nets 50% more points than a mid-range jump shot, but it’s less than 50% tougher to make.

Calvin took 685 shots that were labeled in the play-by-play data as “jumpers”. Well more than any other shot type. I don’t really have a point of reference to compare that to, but it seems like way too many. You obviously can’t count on easy layups all game long, but, to some extent, the payoff is far greater on forced up three pointers than even decent mid-range looks. There are, of course, limits to this statement, but the team shot just under 41% from mid range – to gain that same eFG% on three pointers, one would only need to make only 27.3% of their attempts. And you’d never consider a 27% shooter a three-point threat.

Part of the three point struggles probably had to do with the makeup of the team. A consistent “I’m going to shoot it here” three-point threat didn’t exist on the roster (the team was really limited to opportunistic attempts by Matt DeBoer, Jordan Mast, and Brian Haverdink), so defenses were afforded the luxury of cheating toward the inside to take away Calvin’s mismatch advantages that guys like Tyler Kruis (size) and Tom Snikkers (strength) usually held in one-on-one situations. The result was that many players were probably forced a bit further outside than they would typically like, and jump shots were settled for.

According to the data that I’ve compiled covering all years going back to 1998, Calvin’s historical average effective field goal percentage is .510. It’s probably darn near impossible to shoot that percentage (or better) on jump shots, so it’s imperative that the mid-range look gets minimized. Its only use should be in a game theory sense, to set up the defense for “better” looks on the inside and behind the arc. You have to throw scissors out there every so often, otherwise the defense knows you’re strictly selecting paper and rock, but you only need to employ scissors as much as is necessary to keep the other guessing. It seems clear from the stark eFG% imbalance that Calvin selected the mid-range jumper (scissors, if we extend the metaphor) far too often.

Here are the individualized numbers (sorted by total FGA):


Close
Mid-Range
3-Point
Total
Player
FG
FGA
PCT
FG
FGA
PCT
FG
FGA
PCT
eFG%
FGA
eFG%
Tom Snikkers
71
143
0.497
58
153
0.379
10
50
0.200
0.300
346
0.416
Bryan Powell
52
81
0.642
58
148
0.392
11
41
0.268
0.402
270
0.469
Matt DeBoer
44
70
0.629
74
160
0.463
11
26
0.423
0.635
256
0.525
Tyler Kruis
66
108
0.611
27
74
0.365
0
0
0.000
0.000
182
0.511
Brian Haverdink
8
19
0.421
8
30
0.267
26
80
0.325
0.488
129
0.426
Jordan Mast
11
20
0.550
13
26
0.500
24
52
0.462
0.692
98
0.612
David Rietema
16
38
0.421
7
15
0.467
6
19
0.316
0.474
72
0.444
Mickey DeVries
15
28
0.536
12
25
0.480
3
12
0.250
0.375
65
0.485
Tyler Dykstra
16
26
0.615
8
24
0.333
3
10
0.300
0.450
60
0.475
Mitch Vallie
2
7
0.286
8
14
0.571
11
26
0.423
0.635
47
0.564
Nate Van Eck
14
41
0.341
0
0
0.000
0
0
0.000
0.000
41
0.341
Daniel Stout
0
1
0.000
1
6
0.167
0
0
0.000
0.000
7
0.143
Ryan Nadeau
0
1
0.000
1
1
1.000
2
4
0.500
0.750
6
0.667
Adam DeYoung
0
2
0.000
2
4
0.500
0
0
0.000
0.000
6
0.333
Brent Henry
0
0
0.000
0
4
0.000
0
0
0.000
0.000
4
0.000
T.J. Huizenga
1
2
0.500
1
1
1.000
0
0
0.000
0.000
3
0.667
TOTAL
316
587
0.538
278
685
0.406
107
320
0.334
0.502
1592
0.474


Bullets gleaned from the above data:

  • Tom Snikkers needs to exist close to the block. I know Coach Vande Streek likes his players to move around and dip in and out of the post area, but Snikkers needs to make his living with his inside game. He’s much stronger than most similarly sized players, and that should be the advantage he attacks. He can make the occasional outside shot, but not better than many other guys on the team, and every “settle for” shot he takes is one opportunity where he’s not powering through a weaker defender to get a hoop and/or a foul. I think defenses played off the inexperienced/injured roster and keyed in on Snikkers – clearly Calvin’s go-to guy – and that limited his efficiency across the board. Hopefully things will open up a bit more this year, and hopefully he doesn’t fall into the trap of trying to do too much.
  • Bryan Powell had the best close-range numbers of any player on the team (64%). He ditched the three point shot halfway through the year, which was a good decision, but he replaced it somewhat with a sort-of effective mid-range jump shot. He doesn’t need to ditch that altogether, but his focus on offense should probably be the drive-and-dish and the drive-and-score. Also, the dunk shot should be incorporated more.
  • Matt DeBoer is allowed to shoot from anywhere. When he was open, he was pretty much the best scorer on the floor last year.
  • Tyler Kruis is a monster on the block (61%). He completely eliminated the three pointer from his repertoire this year, but he still took 74 mid-range shots, and made fewer than 37% of them. He needs to stick to the post where he’s an effective scorer and where he’ll draw fouls.
  • Jordan Mast is another guy who’s allowed to shoot from anywhere.
  • David Rietema apparently knows his limitations as a shooter/scorer (he took just about three shots a game as a 22+ minute starter. His shots should come only when open, and only when necessary, though he wasn’t completely horrible on his mid-range and three point attempts.
  • Mickey DeVries probably needs to abandon the three point shot, but showed nice touch on his shot otherwise. He can shoot the ball from some range, but probably shouldn’t be looking to shoot much outside of the painted area.
  • Tyler Dykstra only had limited attempts, but he’s another big, athletic body that will need to get to the basket.
  • Mitch Vallie needs to not get hurt. But he’s another potential “shoot from anywhere” guy. Opportunistically, at least.
  • Same for Jordan Brink.
Of course, it’s far too simplistic to say that everyone should “shoot closer to the basket”, the defense is there to prevent that, but it is important for each player to know their strengths and limitations and realize that sometime even a relatively open shot might not be a good shot.

Want to stay up to date on the happenings at FFTMAG? Follow me on Twitter, “like” us on Facebook, and grab our RSS Feed.

Want a ride to an away game? Check out the Ride Board and post your needs.