Analyzing Sports Streakiness with Texas Tech Professor Alan Reifman........................................................................(See twitter.com/alanreifman for more frequent postings)...................................................................................
Wednesday, November 28, 2012
Wizards Get First Win After 0-12 Start
Earlier today, I put up a poll question asking readers to predict how long the Washington Wizards' season-opening losing streak in the National Basketball Association would continue. It didn't take long for the poll to become moot (and be removed), as the Wizards went and won their first game hours later. With tonight's 84-82 win over the Portland Trail Blazers, the DC squad is now 1-12.
Wednesday, November 21, 2012
Grinnell College's Jack Taylor Scores a Whopping 138 Points in a Single Game!
A basketball player scoring 100 points in a single game, as Wilt Chamberlain did in the NBA 50 years ago and a small number of college players have done as well, is a mind-boggling occurrence. Last night, however, Jack Taylor of Iowa's Grinnell College took things to a new level, scoring an unfathomable 138 points in an NCAA Division III (small-school) win over Faith Baptist Bible, 179-104.
Now, Grinnell's style of play is conducive to the launching of lots of shots. A year ago, another Grinnell player scored 89 in a game. But Taylor scored nearly 50 points more than that 89! According to this 2009 profile of the team's offensive approach:
Grinnell coach David Arseneault employs a scheme known only as “The System.” Sometimes referred to as the anti-Princeton offense, “The System” emphasizes a torrid pace of play where offensive possesions last about 5 seconds and typically end with a three-pointer.
The Grinnell defense is a chaotic, full-court trap that either forces an immediate turnover or degenerates into [an offensive advantage for the other team].
It’s so chaotic, Arseneault has to use platoons, substituting all five players every 45 seconds to keep legs fresh.
Taylor's shooting statistics from last night break down as follows: 25-of-37 on two-point attempts; 27-of-71 on three-point attempts; and 7-of-10 on free throws. He played 36 minutes.
From a hot-hand perspective, the question is whether Taylor actually "got hot" or whether his gigantic point total resulted simply from a mediocre shooting percentage on a huge number of shots (108). I would say it's some of both.
Let's focus on Taylor's three-point attempts, as the distance of the shots is held relatively constant and long-distance shots get the crowd excited. Based on the box score and play-by-play sheet from the game, here are his shooting sequences behind the arc in the first and second half (H = hit, M = miss; shots are grouped into sets of five to simplify viewing):
1st half (9-of-32)
MMMMM HMMMM HMMHM MHHMM MHHHM MHMMM MM
2nd half (18-of-39)
HMMMH HMMMH MHMMM HMHHM MMMMM MHMMH HMHHH HHHH
As shown, for most of the game, Taylor was not hitting more than one or two straight. Any long streaks were of misses. However, to close the game, Taylor hit his last seven three-pointers in a row.
Using this online calculator, I conducted a runs test to detect possible streakiness in Taylor's second-half shooting (I didn't think there would be any in the first half). Each consecutive sequence of hits or misses is considered a "run." The hit to open the second half is one run. The stretch of three straight misses is a second run. The following two straight hits form a third run, etc. Streakiness is characterized by few runs (i.e., staying in long "grooves" of hitting or missing), as compared to what could be generated at random. Here is a screen capture of the results (on which you may click to enlarge; 1 = hit, 0 = miss):
There were 17 runs, which are too many in this context for there to be streakiness. The associated probability (p) level tells us further that there is a 13% chance that a random process (e.g., flipping a coin and looking for runs of heads and tails) could have generated Taylor's second-half sequence of hits and misses from three-point land. Under statistical convention, we would need this probability to be 5% or less (i.e., p < .05).
One thing that I found really interesting about Taylor's final burst of seven straight successful three-pointers is the quickness of the barrage. Below is a screen capture of the play-by-play sheet, to which I've added red arrows to identify Taylor's made threes.
If you click on the graphic, you'll see that Taylor's seven hits all occurred within roughly two minutes of game time (from 3:54 remaining to 1:57). Several times, either a Grinnell steal or a Faith Baptist Bible layup was followed within 3 to 10 seconds later by a Taylor trey.
In my book, I discuss a late-game barrage by Iowa's Justin Johnson, who made six three-pointers in the final 1:56 against Indiana in 2008 to bring his team within two points.
The fact that Grinnell's game against FBB was not close raises another issue, namely the propriety of continuing to launch three-pointers in the closing minutes when leading by roughly 60 points. The desire to let Taylor finish his performance for the ages is understandable. One could argue, however, that it might have been better for Grinnell to start dribbling out the clock once Taylor had surpassed the previous college record of 116 points.
Now, Grinnell's style of play is conducive to the launching of lots of shots. A year ago, another Grinnell player scored 89 in a game. But Taylor scored nearly 50 points more than that 89! According to this 2009 profile of the team's offensive approach:
Grinnell coach David Arseneault employs a scheme known only as “The System.” Sometimes referred to as the anti-Princeton offense, “The System” emphasizes a torrid pace of play where offensive possesions last about 5 seconds and typically end with a three-pointer.
The Grinnell defense is a chaotic, full-court trap that either forces an immediate turnover or degenerates into [an offensive advantage for the other team].
It’s so chaotic, Arseneault has to use platoons, substituting all five players every 45 seconds to keep legs fresh.
Taylor's shooting statistics from last night break down as follows: 25-of-37 on two-point attempts; 27-of-71 on three-point attempts; and 7-of-10 on free throws. He played 36 minutes.
From a hot-hand perspective, the question is whether Taylor actually "got hot" or whether his gigantic point total resulted simply from a mediocre shooting percentage on a huge number of shots (108). I would say it's some of both.
Let's focus on Taylor's three-point attempts, as the distance of the shots is held relatively constant and long-distance shots get the crowd excited. Based on the box score and play-by-play sheet from the game, here are his shooting sequences behind the arc in the first and second half (H = hit, M = miss; shots are grouped into sets of five to simplify viewing):
1st half (9-of-32)
MMMMM HMMMM HMMHM MHHMM MHHHM MHMMM MM
2nd half (18-of-39)
HMMMH HMMMH MHMMM HMHHM MMMMM MHMMH HMHHH HHHH
As shown, for most of the game, Taylor was not hitting more than one or two straight. Any long streaks were of misses. However, to close the game, Taylor hit his last seven three-pointers in a row.
Using this online calculator, I conducted a runs test to detect possible streakiness in Taylor's second-half shooting (I didn't think there would be any in the first half). Each consecutive sequence of hits or misses is considered a "run." The hit to open the second half is one run. The stretch of three straight misses is a second run. The following two straight hits form a third run, etc. Streakiness is characterized by few runs (i.e., staying in long "grooves" of hitting or missing), as compared to what could be generated at random. Here is a screen capture of the results (on which you may click to enlarge; 1 = hit, 0 = miss):
There were 17 runs, which are too many in this context for there to be streakiness. The associated probability (p) level tells us further that there is a 13% chance that a random process (e.g., flipping a coin and looking for runs of heads and tails) could have generated Taylor's second-half sequence of hits and misses from three-point land. Under statistical convention, we would need this probability to be 5% or less (i.e., p < .05).
One thing that I found really interesting about Taylor's final burst of seven straight successful three-pointers is the quickness of the barrage. Below is a screen capture of the play-by-play sheet, to which I've added red arrows to identify Taylor's made threes.
If you click on the graphic, you'll see that Taylor's seven hits all occurred within roughly two minutes of game time (from 3:54 remaining to 1:57). Several times, either a Grinnell steal or a Faith Baptist Bible layup was followed within 3 to 10 seconds later by a Taylor trey.
In my book, I discuss a late-game barrage by Iowa's Justin Johnson, who made six three-pointers in the final 1:56 against Indiana in 2008 to bring his team within two points.
The fact that Grinnell's game against FBB was not close raises another issue, namely the propriety of continuing to launch three-pointers in the closing minutes when leading by roughly 60 points. The desire to let Taylor finish his performance for the ages is understandable. One could argue, however, that it might have been better for Grinnell to start dribbling out the clock once Taylor had surpassed the previous college record of 116 points.
Subscribe to:
Posts (Atom)