Tempo and T2
by Warden11 on Feb 22, 2010 2:26 PM CST
Email
PrintFront Paged because Warden is quickly becoming our resident stat guru and this is some good stuff...Denver
Saturday's game against Colorado reminded everyone why Kansas is the clear #1 team in the nation. Cole Aldrich owned the paint all day. Xavier Henry looked like the guy we all saw in the first half of the season, not just being a solid contributor but the dynamic scorer this team needs at times. Oh, and Tyshawn Taylor was entered back into the starting lineup and proceeded to have one of his best games of the year. A game with that kind of production from those three guys and Kansas will not lose. In reading about the game and Taylor, I came across this at Upon Further Review:
Obviously, if he continues to play this way, KU will probably be unbeatable. But, before you get all googly-eyed about it, keep in mind Taylor plays much better in an up-tempo game – and certainly this was one of those.
Nothing wrong with the quote, according to Statsheet, Kansas had 77 possessions against Colorado. That is the second most possessions they've had in a game this year, so it was definitely a fast game and about 9 possessions more than the national average. What I want to look at today is if the other part of that is true. Do numbers show that Taylor plays better in up-tempo games this year? Now this is where things get a little difficult, how do we determine if Taylor has a good game or a bad game? Because I'm posting this and it is easy to figure, I'm going to use my plus/minus paired with total possessions.
5 games with the most possessions:
| Poss | Plus/Minus |
|---|---|
| 82 | 8 |
| 77 | 16 |
| 76 | 13 |
| 76 | 10 |
| 75 | 12 |
Taylor's had 5 games with a double digit plus/minus and four of those have came in the fastest games of the year. Definitely looks like there is some truth that he plays better as the game speeds up. However, I also have to point out those games are against Alcorn State, Colorado, Tennessee Tech, Hofstra, and California. Not exactly the stiffest competition. Moving on, lets see how he does in the slowest games.
| Possessions | Plus/Minus |
|---|---|
| 65 | 2 |
| 65 | -2 |
| 65 | -3 |
| 62 | 4 |
| 61 | 4 |
The numbers definitely aren't as pretty. But, these numbers are going to be a little lower because they rely on counting stats. Fewer possessions is normally going to equal lower numbers across the board. In an attempt to even that out, I used a multiplier based on KU's average possessions per game (70.6) so I could get a tempo free plus/minus and that is what you see in the graph and table below.
| KenPom | Opponent | Poss | Tempo Free |
|---|---|---|---|
| 347 | Alcorn St | 82 | 6.89 |
| 89 | Colorado | 77 | 14.67 |
| 250 | Tenn Tech | 76 | 12.08 |
| 126 | Hofstra | 76 | 9.29 |
| 23 | California | 75 | 11.30 |
| 138 | Belmont | 75 | 4.71 |
| 79 | Texas Tech | 75 | 1.88 |
| 13 | Missouri | 73 | -2.90 |
| 81 | @Iowa St | 72 | 4.90 |
| 161 | La Salle | 71 | 12.93 |
| 89 | @Colorado | 71 | 2.98 |
| 209 | Radford | 70 | 4.03 |
| 11 | @Kansas St | 70 | 3.03 |
| 27 | @Tennessee | 70 | 0.00 |
| 9 | @Texas | 70 | -1.01 |
| 114 | @UCLA | 69 | -1.02 |
| 74 | Michigan | 68 | 6.23 |
| 68 | Cornell | 68 | -3.11 |
| 15 | Baylor | 67 | 1.05 |
| 81 | Iowa St | 67 | 1.05 |
| 32 | @Temple | 66 | 9.63 |
| 53 | Memphis | 66 | 0.00 |
| 90 | Nebraska | 65 | 2.17 |
| 302 | C Arkansas | 65 | -2.17 |
| 155 | Oakland | 65 | -3.26 |
| 90 | @Nebraska | 62 | 4.55 |
| 34 | @Texas A&M | 61 | 4.63 |
One last graph with my plus/minus taken out and offensive rating put in its place. Graphs are very similar.
*note- I did not include the Baylor game in this graph, Taylor played 15 minutes and finished with an offensive rating of 269 while going 0-3 from the field with no rebounds.
Definitely appears as though Taylor does play better as the game speeds up. Though a +4 in the two slowest games or.f the year isn't bad eithe
33 comments
|
3 recs |
Do you like this story?
Comments
good stuff, rec'd
It might be good to consider the strength of the opponent (as you mention) numerically along with the possessions and +/-. While a multi-variable regression including a variable like “Team’s KenPom rating” would be super duper, I don’t begrudge anyone for not wanting to open the regression can of worms. Maybe just listing the KenPom of the opponent next to the Best 5/Worst 5 would give some perspective.
Maybe better teams slow us down, and that causes TT to be less effective (either b/c they’re better or b/c they’re slow), which would be an important piece of info (for us armchair coaches).
Zapp Brannigan/Dayton Moore quote of the day: "...And like all my plans, it's so simple an idiot could have devised it!"
by SagehenMacGyver47 on Feb 22, 2010 3:00 PM CST reply actions
Good thinking, I should have listed either KenPom or the RPI in the table that shows all of them.
I’ll do that and then edit.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball. I’ll tell you what I do. I stare out the window and wait for spring.
looks good - at first glance i didn't see a trend toward good opponent = slow pace
I did a quick check with a regression (opponent’s KenPom vs Possessions), and had a negligible R^2 (0.107) which refutes my earlier idea that maybe good teams slowed KU down. The slope indicated that an increase in KenPom of 50 places (i.e., #10 to #60) would add only 1 extra possession to the game.
Seems to make sense: worse teams lead to faster-tempo games, but not by a large amount and not with any easily-detectable significance.
Zapp Brannigan/Dayton Moore quote of the day: "[my most important sabermetric stats are] runs scored and runs driven in"
by SagehenMacGyver47 on Feb 22, 2010 5:15 PM CST up reply actions
ok, further review
hopefully not bogging people down here with this.
Regression with “KenPom ranking” and “# of Possessions” as variables determining “TT’s Tempo-Free +/-”. Any regression results taken with a grain of salt of course, and the R^2 was only 0.21, so this is far from “figuring out” what makes TT tick, but it should help isolate the effect of possessions vs quality of opponent.
The KenPom ranking had a negligible coefficient in the hypothesized direction (50-place increase led to 0.05 increase in TT’s +/-), though without any statistical significance. # of Possessions, however, had a larger coefficient, with significance; an increase of 1 possession lead to an increase 0.5 in his +/- (Tempo Free), if all else were equal.
Usual caveats apply: small sample size, regressions are fidgety at best, etc., but the relative strength of the results was striking enough that I had to mention them.
Zapp Brannigan/Dayton Moore quote of the day: "[my most important sabermetric stats are] runs scored and runs driven in"
by SagehenMacGyver47 on Feb 22, 2010 5:33 PM CST up reply actions 1 recs
Definitely not boggin anything down here,
the more discussion the better.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball. I’ll tell you what I do. I stare out the window and wait for spring.
I looked if to see if he played better against the weaker oppenents
But there was almost no correlation. Slope was almost zero, and the R^2 value was 0.02 when I used Kenpom numeric ranking vs tempo free stats. I know I shoud have used their Pythagorean expectation numbers not the ordinal numbers were easier to cut and paste, and the results did not warrant more exploration.
by bt01 on Feb 23, 2010 12:25 AM CST up reply actions 1 recs
I'm just rec'ing all this stuff because I have almost no clue what they're talking about...
But it sounds scientificky and junk.
I do remember Rsquared values, however, and know that 0.02 is incredibly small- almost no correlation at all…right?
"Here are our top priorities: recruit, beat Missouri, recruit, win the North, recruit, win the Big 12, and in most cases if you win the Big 12 then you're playing for a National Championship. And then we're going to recruit."
by KennyGregoryRockThaCradle on Feb 24, 2010 10:38 AM CST up reply actions
Yep, that's what it is telling us.
Here’s the scatter plot.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball. I’ll tell you what I do. I stare out the window and wait for spring.
As Austin Powers would say
NERD ALERT!
by Andrew Clark on Feb 24, 2010 10:51 AM CST up reply actions
eh, I had that done when I typed the post.
Didn’t include it because it didn’t show any new information.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball. I’ll tell you what I do. I stare out the window and wait for spring.
As a Math Teacher
I take that as a badge of honor. Someone you call a nerd when you are young is someone you call “boss” when you get older.
Did I mention I took Calculus in HS with Ken Pomeroy?
by bt01 on Feb 25, 2010 9:48 PM CST up reply actions 1 recs
Awesome.
This is really great stuff you just can’t find anywhere else.
On a slightly related note, does anybody else find Upon Further Review to be kinda amateurish for supposedly having the support of a major US newspaper? I think it’s a combination of the layout, which kinda sucks, and the fact that they respond in the comments to any snide comment that’s made their way, instead of just ignoring it. And like Warden points out here, they rely too much upon counting stats, IMO.
Basically, RCT >>> UFR
"Here are our top priorities: recruit, beat Missouri, recruit, win the North, recruit, win the Big 12, and in most cases if you win the Big 12 then you're playing for a National Championship. And then we're going to recruit."
by KennyGregoryRockThaCradle on Feb 22, 2010 3:17 PM CST reply actions
You're not alone
sometime I read their stuff because it’s usually a little more in-depth than what J. Brady gives us but yeah, the engine for all of the Star’s blogs really sucks.
Me no likey T2.
agreed
I appreciate UFR/Martin’s (?) numeric objectivity, but it is just anothr place that RCT kicks butt on the “Pro’s”.
So does Taylor play better as a result of the tempo...
or does Taylor force the tempo more? Some guards would appear to have the ability both through perimeter pressure and pushing the offense to increase the pace of a game singlehandedly. Our very own Jacque Vaughn comes to mind. Does Tyshawn have that ability when he’s on?
52 Conference Championships!! Holy Hell...Good Luck with That!!
anecdotally, I think yes, he speeds up the game
From what I remember of Tyshawn’s most memorable moments (how’s that for repetitive redundancy?), he was flying down the court, not because the opposing team had taken a quick shot that resulted in a long rebound, but because he’s just normally starting out in 3rd or 4th gear.
Maybe another interesting column would be average possessions per game for our opponents, to see if KU sped them up, or if they are a naturally fast team.
Actually thought about something to do with possessions last week
and wondered if the team seemed to play better with more or fewer possessions. Got into the numbers and became bored. Looking at the list above, it appears that Kansas more or less plays at the opponents preferred pace (that’s eyeballing the list), Saturday against Colorado is a major exception to that.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball. I’ll tell you what I do. I stare out the window and wait for spring.
yeah, that would be interesting
who deviates more from their average, KU or the opponent?
Zapp Brannigan/Dayton Moore quote of the day: "[my most important sabermetric stats are] runs scored and runs driven in"
by SagehenMacGyver47 on Feb 22, 2010 5:17 PM CST up reply actions
I seem to recall
Jaybate, LJW blogger, commenting that this was a tenant of Coach Self/Sutton/Iba-Ball
sort of like beating them at their own game?
hmm…that Jaybate guy
sure wish we had someone with that kind of insight around here
Zapp Brannigan/Dayton Moore quote of the day: "[my most important sabermetric stats are] runs scored and runs driven in"
by SagehenMacGyver47 on Feb 23, 2010 11:37 AM CST up reply actions
I remember it, too
It was in one of the longer posts (which ones aren’t?), but yeah, it’s take what they give you until they’re not able to stop you from doing what you want. Frustrate them by beating them at their own game until they make mistakes and give you even more.
He seems to thrive on the secondary break.
Slashing when the opponents have gotten down court, but have not fully set up their defense.
The factor that would seem significant to me,
and a place where having a smaller sample size would be a downfall, has to do with the change in TT’s play thru the year – from starting, to being benched/challenged to play better, to starting and controlling games like he really hasn’t done all year.
The other similar factor is the development of McM and even Cole from mid-Jan, and now redevelopment of X. I realize there is individual play, but there is no denying the impact of having teammates playing quality minutes.
Ultimately, I think what we have here is a young kid that’s still inconsistent at the college level. There doesn’t seem to be one specific on-court factor that explains his ups and downs. Query whether there are any off-court issues that might be messing with him- family issues, girl issues, football team issues, who knows.
What we do know is that he’s played pretty well since getting back into the starting lineup, and should be a key going down the stretch.
Tyshawn Taylor, you are an enigma wrapped in a riddle surrounded by a mystery.
"Here are our top priorities: recruit, beat Missouri, recruit, win the North, recruit, win the Big 12, and in most cases if you win the Big 12 then you're playing for a National Championship. And then we're going to recruit."
by KennyGregoryRockThaCradle on Feb 24, 2010 10:42 AM CST reply actions
I should just copy and paste that and put it in as my conclusion.
Though over the weekend I’m going to add last year’s game to increase the sample size.
Glad I came, just wish I hadn't stayed so long.
People ask me what I do in winter when there’s no baseball. I’ll tell you what I do. I stare out the window and wait for spring.
Can you add a factor
to diminish the significance of the game/results the older the game is?
not speaking for Warden here
but the first problem that comes to mind for that is that it’s pretty subjective — how much to you discount the past? 99% as much as the next game? 90%? 75%? But as long as it’s understood that the factor (and results) is subjective, it still may be interesting.
Another way to approach it would be to add a time variable (i.e., days into the season) into a multi-variable regression and then you can simultaneously see if time played a factor and control for the effect of time on the other variables.
Zapp Brannigan/Dayton Moore quote of the day: "[my most important sabermetric stats are] runs scored and runs driven in"
by SagehenMacGyver47 on Feb 25, 2010 11:50 AM CST up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
