Regression From The Mean


Walkon
Jan 07, 2019
[]
[]
4 Comments

I enjoy Las Vegas. I've probably been to Sin City 20 plus times over the last 30 years. I'm primarily a blackjack guy, but I'm also always oddly fascinated by roulette. I don't gamble much on the wheel, but it's hard for me to walk by an active one without a quick glance to see if my lucky numbers are populating the "recent spin result" tote boards. Invariably you will come across one with six or seven black or red winning numbers in a row. Yet despite an extended run of one color or the other, the odds on the next spin never change, and so you know that eventually red or black will find its way back to even.

I keep thinking about that when it comes to Illinois basketball over the past season and a half. Obviously basketball games depend on a human element and are not 50/50 propositions, but individual possessions - offensively and defensively - aren't statistically that far off from a single roulette spin. With that in mind, I have to ask myself how many more times that ball can come up black when Brad Underwood has all his chips on red.

After Illinois lost to Northwestern today in yet another heartbreaker, Brad Underwood made a point to say that he's been coaching for 31 years and he believes that games like these even out over time. You know - I'm starting to wonder.

So if you would, please accompany me on a painful trip back over the past 13 months (come on - it will be cathartic)...

+ Illinois @ Northwestern 2017: Game tied 63-63 with 45 seconds left. Illinois turns it over after a Leron Black offensive foul and goes on to lose in OT 72-68.

+ Illinois vs Maryland 2017: Illinois leads Maryland 74-72 with 4 seconds left. All they have to do to win the game is successfully in bound the ball. Da'Monte Williams throws it away and the Terps tie the game in regulation as time expires. Game goes to OT and Maryland hits a single free throw with :01 on the clock to win.

+Illinois vs New Mexico State 2018: Game tied 69-69 with 34 seconds left. Illinois ball under it's own basket. Aaron Jordan turns the ball over on the inbounds play and NMSU gets a breakaway lay-up and one. Illinois loses 74-69 in the United Center.

+ Illinois @ Nebraska 2018: Mike Finke converts a four point play to put Illinois up 64-62 with 6 seconds left. Illinois loses 65-64 as James Palmer nails a three for the Huskers as time expires.

+ Illinois vs Georgetown 2018: Illinois trails 81-80 with one minute remaining. They need a defensive stop or to at least hold the Hoyas to two points to maintain the one possession game status. Illinois gives up a three point play and goes on to an 88-80 loss.

+ Illinois vs Gonzaga @ Maui 2018: Illinois comes back from down 10 to pull within two against the Zags at 80-78. A three pointer from Trent Frazier is in the air with 16 seconds left. Air ball. Illinois loses 84-78.

+ Illinois @ Notre Dame 2018: Another three pointer is in the air from Frazier with Illinois down 76-74 and the clock hitting :00. Halfway down...and out. Loss.

+ Illinois vs Florida Atlantic 2018: Illinois ties it on an Ayo Dosunmu three to force OT. Illinois misses three of four free throws in the final minute of overtime and also miss a game winning three attempt from DMW as OT expires. The worst home loss in a decade.

+ Illinois @ Northwestern 2019: Today. Illinois enters the final minute of play up one at 66-65. Over that final minute, the Illini commit a shot clock violation, give up a three to Northwestern off an offensive rebound, and miss a game winning three point attempt. They lose yet again to fall to 4-11 overall and 0-4 in the Big Ten.

Even when Illinois has been successful in last possession chances they've ultimately failed to close the deal. In addition to the FAU loss noted above, there was Iowa last January at the SFC. Trent Frazier capped a crazy comeback with a game tying three to force OT, but the defense was not up to the task - giving up 14 points in overtime in a 104-97 loss. Did I mention that Brad Underwood is 0-4 in overtime games at Illinois?

The only winner out of 10 "final possession" opportunity games was last December at the SFC when DMW hit a three with :17 left to help Illinois hold on to beat Grand Canyon 72-68.

That is 10 games over the past two seasons which came down to a final possession or overtime - again almost a 50/50 proposition on each - and 9 of 10 times the Illini rolled snake eyes. You keep thinking that Illinois is due for a positive regression to the mean - yet they continue to defy statistical probability.

KenPom even has a "Luck" stat. This is an actual thing that exists. It's basically a measure of the deviation from a team's expected record and their actual record. Even more simply - if a team is involved in a large sample size of close games, they really shouldn't win (or lose) the majority of them.

Illinois ranks 352 out of 353 teams in "luck" this season and ranked 343 out of 353 last season. How in the actual what is that even possible?

You can choose to focus your blame on the coaching staff for not putting the players in the best position to win or you can lament players not stepping up to make plays. Take your pick - there is plenty of blame to go around. The answer is likely somewhere in the middle, but nonetheless at some point simple math has to come into play right? Regression to the mean is real right? RIGHT?

I think we all knew this season was going to be a struggle, but man the journey has been unusually cruel. Brad Underwood said after the game today that he had a locker room full of guys who were "hurting", but he viewed this as a positive because it showed him a team who cares and who have "completely bought in". I am happy to hear that, and I genuinely hope he's right, because it sure seems like continually losing games this way can become habit forming - statistics be damned.

KOFI!

So today wasn't all bad. I'm sure Robert is chomping at the bit to write an LLUOI post on the Kofi Cockburn commitment, but I would be remiss not to at least acknowledge the huge (pun intended) recruiting win from this afternoon.

In our post-Indiana game Back/Forth, Robert and I presciently discussed how the potential addition of a "GIGANTIC HUMAN" such as Cockburn would impact next year's roster and lineup. And make no mistake - at 6' 10" and 290 pounds Cockburn is indeed a GIGANTIC HUMAN. Yet beyond the "on the court" impact, this commitment could also go a long way in changing the entire narrative of the program moving forward.

Underwood and his staff have taken heavy criticism on the recruiting front of late, but now suddenly, Underwood has his second top 35 level recruit in as many years. And even though we should note that no LOI's have been signed, the Illini frontcourt rotation is now basically set for the next two seasons with Giorgi, Samba Kane, Antwan January, and now Kofi.

When you're trying to sell a program in the midst of it's worst season in 40 years - pulling a highly ranked center from the East Coast almost out of nowhere is nothing short of miraculous. This commitment lends immediate credibility to other ongoing recruiting efforts as Underwood continues to look to complete his 2019 class with some much needed wing athleticism and scoring.

So welcome GIGANTIC HUMAN!

Comments

illiniranger on January 07 @ 10:30 AM CST

roulette is a game of chance. basketball is a game of skill (including the coaches skill). we are not very skilled, so i don't know why Underwood thinks this will magically even out over time.

Walkon on January 07 @ 11:04 AM CST

Yes, of course, on the skill vs chance thing - but in observing 10 unique possessions - it’s statistically improbable for any team to fail 9 of 10 times.

The top half of the “luck” stat on KenPom includes quite a few lousy teams. Point being that even bad teams will usually win a fair share of close games.

FlyinIllini22 on January 07 @ 06:47 PM CST

Coach Henson had an incredible OT record he probably won over 80%. I know he went several games in a row winning those games.

Bear8287 on January 08 @ 01:34 PM CST

Because I can't help myself :-D ...

Yet despite an extended run of one color or the other, the odds on the next spin never change

Yes, this is true.

You keep thinking that Illinois is due for a positive regression to the mean

Okay, here's where your reasoning perhaps starts to get a bit wobbly.

If in fact it's really a 50/50 proposition, the expectation that they're eventually going to win one of these games is reasonable.

The expectation though that somehow it's more likely that they're going to win because of all of the losses that they've incurred, or that they'll go on some type of winning streak to "even it out" is not. Just like that roulette wheel, the odds on the next spin don't change.

So how does this regressing to mean stuff work? Think Law of Large Numbers. Let's take the close games 1-9 record. At the moment the Illini have won 10%, which is pretty far from the expected mean of 50%. Now suppose that the Illini play 90 more close games and as expected win 45 of those. Over 100 games they would've now won 46% (46-54). Suppose that they play 900 more and win an expected 450. They would be 496-504 and would have won 49.6% of those games, an astonishing improvement from 10% and this can all happen without any miraculous improvement by continuing to just have a 50/50 chance of winning any close game.

Now the odds of there being streaks of consecutive wins or losses over say 1000 games is pretty high, but the odds of those streaks being wins or losses is the same.

So, if it's really a 50/50 proposition then it's reasonable to expect that the Illini should start winning these types of games at a 50% clip. If it turns out that it's not really a 50/50 thing then all bets are off. :-D

Speak Your Mind

Please login or register to post comments on the IlliniBoard.