A Video Charter’s Guide To The Final Four | Stat Geek Idol

posted in NCAA Basketball, NCAA Tournament, Stat Geek Idol

This is a Final Four submission in our inaugural Stat Geek Idol contest. It was conceived of and written by Jordan Sperber of Hoop Vision (@HoopVision68). For more from Jordan check out his previous round posts.
Sweet 16 entryVideo Charting Wisconsin’s Swing Offense
Round of 64 entryQuantity Over Quality: Getting More Whacks At The Piñata

Last week, I attempted to quantify Wisconsin’s Swing Offense through the use of video charting. With the Final Four, I set out on a similar goal for Louisville, Ohio State, Kentucky, and Kansas.

To analyze each offense I watched each offensive possession from the Elite Eight of the Final Four teams in slow motion and charted categories such as: number of passes, shot location, defender proximity, number of dribbles the shooter takes, time of possession, number of players crashing the boards, and more. When it was all said and done, I was about 3% closer to becoming a video charting expert.

Can We Trust This Data?

Video charting is a very long process and therefore a large sample size is hard to get. In my case, I charted four games worth of information. Think about a statistic like passes per play. This is not a conventional metric and thus we do not really know how much a statistic like this would vary from game to game.

To try and explore the variation of passes per play, I used two much more conventional statistics: three pointers attempted per game and three pointers made per game. I randomly picked Ohio State as my team to examine. On the season, Ohio State averages about 5 threes per game and about 15 three point attempts per game. I wanted to see how consistent these two stats were by looking at them on a game by game basis. The results were as follows:

OSU was NOT within plus or minus 25% of their average 3PA in 5/38 games (13%)
OSU was NOT within plus or minus 25% of their average 3PM in 16/38 games (42%)

The point I’m trying to make here is it appears that the “process” is much more consistent than the result. The above shows evidence that passes per play in one game should be fairly consistent. However, the results from these passes are going to be much more likely to differ due to random variation.

For what it’s worth, let’s say there is in fact a 13% chance that any one game I video charted was a bad representation of the team. Therefore, there’s about a 3 in 10,000 chance (.13^4) that everything I say in the rest of this article is wrong. That’s just about the same probability of South Dakota State winning the National Championship according to Ken Pomeroy’s pre-tournament log 5 analysis, so that makes me feel much better.

A Word About Ohio State

In the Elite Eight, Syracuse’s 2-3 zone successfully altered the Ohio State data. The bad news is that it is not a good idea to compare OSU’s offense to the other three teams who played against man.

The good news, however, is that OSU’s stats do show evidence of what we would intuitively expect would happen against a zone. There will be more on this later, but I also put Wisconsin data from my last piece in two graphs below that actually show that OSU (and presumably most teams) against zone is very similar to Wisconsin (and presumably just about no other teams) against man.

Comparing Style Of Play

(click to enlarge)

First, let’s look at the two scatter plots on the left of this image. The y-axis for both plots is time of possession. This shows us that Kansas played the fastest in the Elite Eight, followed by Kentucky and Louisville. All three teams, however, were very close (average plays were between 11.4 and 13.6 seconds). Ohio State was clearly affected by the 2-3 zone of Syracuse. In fact, OSU has played at the fastest tempo of any of these teams this year (69th in the country). I put Wisconsin’s numbers from the Vanderbilt game in the graph to show just how slow Syracuse forced OSU’s offense to play.

Looking at the top scatter plot, you see the relationship between time of possession and passes. The r-squared value tells us that 79% of the number of passes in a play can be explained by the amount of seconds in that play. Teams below the line of best fit relied relatively more on passing, while teams above the line of best fit relied relatively more on dribbling. Here the results came out as I would have expected. Louisville’s Peyton Siva and Russ Smith tend to over dribble, while Ohio State definitely relied on quick passes against the zone.

The second scatter plot not surprisingly has a strong correlation. It shows the relationship between average time in between a potential scoring opportunity and time of possession. It is hard to define a potential scoring opportunity, but I tracked any time a player with the ball did something with the intent to score. Louisville being above the line of best fit shows that they look to score more often then would be expected by their time of possession. This seems to imply that Louisville is trying to play faster and could be predictive of a faster pace in the future.

The two bar graphs simply show the average number of dribbles taken by the player who shoots and the number of players who crashed the offensive glass. The latter was pretty similar for all four teams. Louisville and OSU were just below two players and Kansas and Kentucky were just above two players. However, the average number of dribbles again points out the isolation style of
Louisville’s guards.

What I Noticed While Charting

Watching every offensive play in slow motion for four teams doesn’t just help you with statistical analysis. Below are screen caps depicting the main basketball observations I took from the four teams:

(click to enlarge)

Peyton Siva likes to jump. Fundamental basketball says don’t leave your feet to pass. In the Louisville-Florida game, I noticed Siva jumping on his drives and therefore creating high level of difficulty passes. Interestingly enough, Siva had an assist:turnover of 8:1 in the Florida game, so his jumping problem didn’t hurt him in the Elite Eight. On the year, however, Siva struggles with a VERY high Turnover Rate (TORate) of 29.4%.

Doron Lamb and Darius Miller use off ball screens well. The photo above shows Lamb coming off a double screen on a baseline out of bounds play. Both Lamb and Miller are in the top 200 in effective field goal percentage (eFG%) and move extremely well without the ball. When thinking of Kentucky it is easy to only remember the dribble drive, but those responsibilities are left to Jones, Kidd-Gilchrist, and Teague.

Thomas Robinson effectively uses the post pin. Robinson’s positioning in the post was extremely impressive against UNC. He does a great job of thinking a step ahead, temporarily giving himself poor positioning in order to get an easy basket later in the play. Kansas’s offense is clearly designed to take advantage of this, notice how the lob pass is wide open with all four Kansas players drawing their defenders away from the basket. In fact, Self talks about this post positioning in his High Low Motion Offense DVD. Robinson is shooting 51% from two.

Deshaun Thomas might have court vision. I was impressed with Thomas’s interior passing against the zone. In this case, the numbers seem to say otherwise. Thomas had just one assist on the game and has an anemic 5.9% Assist Rate on the season. Still, multiple times in the Ohio State-Syracuse game Thomas delivered timely passes inside to Jared Sullinger and Amir Williams who either missed the lay up or got fouled. Thomas is no Greg Monroe, but his vision did impress me.

The Players: Louisville Vs. Kentucky

(click to enlarge)

The Louisville backcourt is led by Peyton Siva and Russ Smith. Both players have efficiency problems. Against Florida, 16 of the 23 shots the duo took were fully contested. These two tend to shoot off of pick and rolls and take a lot of dribbles. On average, Siva dribbled 7.9 times before shooting and Smith dribbled 3.5 times before shooting. On the other hand, Kyle Kuric took a total of ONE dribble before all nine shots he took. Kuric got open looks from three (especially the corner), but did struggle with results against the Gators.

In the frontcourt, Behanan and Dieng took a lot of contested hook shots and lay ups like traditional big men. Dieng struggled finishing against Florida while Behanan had a very nice game. Dieng especially sets a lot of ball screens for Siva, but as you can see only took one shot as a roll man.

For Kentucky, Anthony Davis put up quite an impressive line against Baylor. On his 12 shots he took a total of just three dribbles and consistently took good shots. Just three of Davis’ 12 shots were highly contested and all three of those were lay ups. Apart from an occasional Teague isolation, Kentucky’s starters did not over dribble against Baylor. Much has been made about Davis and the Wildcats’ defense, but their offense is probably even more impressive.

The Players: Kansas Vs. Ohio State

(click to enlarge)

The thing that stands out when looking at Kansas and Ohio State is point guard play. In Tyshawn Taylor you have a high usage, scoring point guard. In Aaron Craft you have the opposite of that. Tyshawn Taylor took shots of all types and from many different locations against UNC. The one common thread for Taylor’s shots was that the great majority of them were contested. However, Taylor’s shooting numbers have still been very good on the year.

You can tell just how hard it is to get good shots up against Syracuse’s zone. Smith, Buford, Sullinger, and Thomas all took plenty of contested shots in the Elite Eight. Sullinger’s shot attempts are low largely due to first half foul trouble. Both Sullinger and Thomas never took more than one dribble before taking a shot. In the backcourt, Buford and Smith both shot plenty of spot up threes against the Orange.

For Kansas, Thomas Robinson used strong post position as well as rolling to the hoop off his picks to have a nice game against the Tar Heels. On the other hand, Jeff Withey didn’t take a dribble and didn’t miss a shot, so there’s that. Elijah Johnson tended to operate in isolation and pick and roll situations, while Travis Releford was more of a spot up shooter.

Conclusion

My video charting only covered the offenses of the Final Four teams, but it was clear Kentucky, Ohio State, and Kansas are superior to Louisville. Fortunately for Pitino’s squad, they have the number one rated defense in the country. The Kentucky-Louisville rivalry is getting all the hype, but the OSU-Kansas game should be a phenomenal game. Video charting is a long process that requires a lot of patience, but is a great way to get to know a basketball team.

Remember, for more from Jordan check out his posts from previous rounds.
Sweet 16 entryVideo Charting Wisconsin’s Swing Offense
Round of 64 entryQuantity Over Quality: Getting More Whacks At The Piñata
And for more from the other contestants, explore the Stat Geek Idol category on the blog!