The Bias Chart

This is the fourth in a weekly-ish series of two-game recaps of the 2021-22 Tennessee men’s basketball season.

December 4: Tennessee 69, Colorado 54 (6-1)
December 7: Texas Tech 57, Tennessee 52 (OT) (6-2)

Firstly, thanks for the kind words on yesterday’s post. If you have my brain, there are certain ideas you feel you are legally required to get out of your system. That was one, and it meant more to me than nearly every other piece of writing I have published has.

In the process of waiting to hit the Publish button, I spotted this on Twitter dot com:

And again, if you have my brain, this instantly became the next post about Tennessee’s men’s basketball team.

After spending Sunday evening catching up on Saturday’s Colorado fixture, which I didn’t see a second of in real time, I wish I’d been able to do the same with Tuesday’s Texas Tech affair. Games like that are so bad that I spend an alarming amount of self-indulgent time looking up stats like this:

Or attempting to see the last time a pair of high-major basketball teams were this incompetent on the court offensively and realizing it is possibly the single worst game between two good teams in recent memory.

The Act of God Game is something that overshadows everything that came before it. The working title of this post was the Milhouse Games, because in one (Colorado) just about everything went as correctly as it reasonably could have and in the other (Texas Tech) it was a pure disaster on the level of Kirk van Houten (Milhouse’s father) revealing his racecar bed. But because Act of God is more recent, it completely, totally overshadows everything that came before.

That’s #15 on the chart, Recency Bias. All of us are guilty of this, but it feels perhaps the most prevalent it’s ever been. A hyper-reactionary society addled by apps and satiated by the latest seven-second videos is going to be this way, and I include myself in it. After Colorado, you could have easily convinced me Tennessee was well on their way to being one of the eight best teams in the sport; after Texas Tech, I am worried I won’t want to watch another Tennessee game for a month.

The Bias Chart is helpful in many ways, but it somehow seems like any and all of its 20 labeled biases apply to this season just eight games in, and perhaps how I have personally covered the team thus far. Let’s see just how many ways I can put myself in a mental pretzel.

1. Anchoring bias.

People are over-reliant on the first piece of information they hear.

Well, yeah, it came first, so of course I am. Tennessee began this season by setting a program record for made threes in a game while taking the second-most three-point attempts in a single game, too. Prior to Act of God, they hadn’t breached 30 in any game since. Tennessee is taking more threes this season, but it’s in a sense of attempting to get 24-28 threes off each game rather than 40 like Rick Barnes hinted at. That’s fine, but when the first memory you have is of the team taking 46 against Lenoir-Rhyne and 40 against Tennessee-Martin, it will naturally feel a little disappointing. Then again, maybe we would like to never see a three-pointer after Act of God.

2. Availability heuristic.

People overestimate the importance of information that is available to them.

This is important because studies in the NBA – which has a larger sample size – show that offensive rating doesn’t stabilize until 13 games in (624 minutes of basketball), which converts to about 15.6 games of college basketball. Tennessee’s offense, thanks to two turd performances and a third wobbly one against Colorado, looks like another bad offense. Yet the performances against North Carolina and ETSU would suggest a very good offense. What if we are overestimating eight games’ worth of knowledge? I certainly have been.

3. Bandwagon effect.

The probability of one person adopting a belief increases based on the number of people who hold that belief.

Groupthink, in other words. I am naturally not one to stray too far from what smarter writers say about basketball; that has left me in a place where I don’t want to say anything out of line. Also this is about Tennessee basketball’s stinky offense again sorry.

4. Blind-spot bias.

Failing to recognize your own cognitive biases is a bias in itself.

This is about the threes, which I have pushed for for years without realizing until this summer that three-point attempt rate has almost no correlation at all to a better offense. The lack of mid-range attempts does, but replacing them with threes doesn’t immediately give you a better offense. Better, more consistent shooting regardless of the spot on the floor gives you a better offense. I do think Tennessee should take a lot of threes, because based on the careers of various roster members, they have more shooting options than the average high-major team. But as we’ve seen, that alone isn’t everything.

5. Choice-supportive bias.

When you choose something, you tend to feel positive about it, even if that choice has flaws.

I call this one “writing about the Tennessee men’s basketball program for five seasons.”

6. Clustering illusion.

This is the tendency to see patterns in random events. It is key to various gambling fallacies.

Also known as the “hot hand fallacy.” The hot hand is partially a truth in the NBA, but only for certain players; think the Damian Lillards and Stephen Currys of the world. It is all randomness otherwise. This is how the same team that did this against top 20 competition on a neutral floor:

Can be the same team that did this barely two weeks ago against top 50 competition on a neutral floor.

If you just watched the North Carolina game, you would believe Tennessee’s offense is one of the ten best in America. If you watched Act of God, you would believe Tennessee and Virginia will be competing for the 2021-22 Sickos Team of the Year Award. The truth most likely lies somewhwere in between.

7. Confirmation bias.

We tend to only listen to information that confirms our preconceptions.

The preconception any non-biased observer has of Tennessee basketball is that it’s a program that wins a lot of games by way of tough defense and, uh, attempting to not get out-shot rather than attempting to out-shoot. Through seven games, Tennessee had done a terrific job of getting rid of preconceptions, eschewing mid-range jumpers at a significant rate for more threes and slightly more attempts at the rim. All it takes is one Act of God to erase all of the good vibes that was producing, even when the Act of God statistically shows you got 90.1% of your shot attempts at the rim or from three.

8. Conservatism bias.

When people favor prior evidence over new evidence or information that has emerged.

go to a game, wait for the first dry streak from three, listen to the people seated around you, then get back to me

9. Information bias.

The tendency to seek information when it does not affect action. More information is not always better. With less information, people can often make more accurate predictions.

As the Information Guru here this bias personally offends me, which is why I am obviously very much a sinner in its eyes. But, like…this one seems as if it doesn’t really apply to Tennessee? All of the information heading into this weekend pretty much played out: a great rim defense forced a bunch of bad attempts at the rim, didn’t give up many points inside the perimeter in general, and made life crazy difficult on both opponents. They’re clearly very good, and I don’t know what good waiting for two Quadrant 4 opponents will do.

But if you just look at that bolded part, it makes sense. If I simply did not know that Tennessee played Texas Tech in a college basketball game, I would be this guy.

Alas, because I Investigated, I am this guy.

10. Ostrich effect.

The decision to ignore dangerous or negative information by burying one’s head in the sand, like an ostrich.

NEVER HEARD OF IT!

11. Outcome bias.

Judging a decision based on the outcome – rather than how exactly the decision was made in the moment.

Here are four screamingly wide-open shot attempts.

The shooters attempting these four shots, prior to last night, were a combined 38-for-100 (38%, obviously) on threes this season. Considering that Synergy’s numbers have consistently shown wide-open catch-and-shoot threes to be worth about 3% more than the average three-point attempt, we could take a reasonable guess that these players would be expected to hit 1.64 of these four attempts. The odds of all four missing are 12%, just in this small sample size. All four shots have no defender within five feet. All three shooters have made at least three threes in a game this season.

None of the shots went in. In fact, Tennessee did not make a single open catch-and-shoot three, per Synergy. (They credit Texas Tech as guarding 26 of 33, which seems really generous based on how it looked to me.) That doesn’t mean that these shots shouldn’t have happened. They were right to shoot the threes, because a lot of them were open.

Just like they were right to attempt all of these against Colorado:

Despite none of those going in, either.

12. Overconfidence.

Let’s not do this one plea-

Some of us are too confident about our abilities, and this causes us to take greater risks in our daily lives.

Fine.

I am confident that Tennessee has a roster of good shooters. Of the ten members of the current rotation (yes, ten), nine have hit a three in a college uniform and six have hit at least one per game this season. That’s pretty darn good. No other SEC school currently has more players with 6+ made threes. Tennessee has had three bad performances and five wonderful ones from downtown.

And yet: do you think they might’ve been a little too confident in the concept of shooting over the opponent? Colorado was much happier to let Tennessee take threes than to allow them to keep murdering them at the rim. Texas Tech was great at the rim, but Tennessee barely put any pressure on it in the final 10 minutes of regulation and, per StatBroadcast, recorded zero points in the paint from the 11:49 mark of the second half to the 2:01 mark of overtime. That’s unacceptable even if you are killing it from three. If Tennessee does want that confidence to not feel misplaced, this cannot happen again.

13. Placebo effect.

When simply believing that something will have a certain effect on you causes it to have that effect.

14. Pro-innovation bias.

When a proponent of an innovation tends to overvalue its usefulness and undervalue its limitations.

Time for you to take a wild guess as to what this is about.

In my own self-importance, I’ve violated the “don’t send excessively mean tweets about individual players” rule of mine a few times this season. Almost all of them have been about Uros Plavsic, the 7-footer that I instantly groan whenever I see him enter the court and count the seconds until he exits. The sad thing is that many, many Tennessee basketball fans are in the same boat. And truth be told, they probably should be. Those numbers for his career hold true. For 2021-22 specifically, adjusted for opponents and for 3PT luck, it instantly makes me feel like an ass.

The team is effectively just as good with him on the court, with almost no serious change in shot selection, as they are when he’s off. The defense is significantly worse, but Plavsic is genuinely useful as a screener and does not usurp many possessions of his own. I think that he is an openly bad rebounder for his size – of the 62 7+ footers playing at least 8 minutes a game in CBB, he ranks 59th in defensive rebounding percentage – but I’m going to be real here for a second. If multiple Big Six schools have watched Plavsic play and offered him a scholarship, even if one of those schools was coached by Bobby Hurley (a bad basketball coach!), there’s talent to be had.

Re-running the numbers for Brandon Huntley-Hatfield based on the second half of this tweet made me feel worse.

BHH’s numbers are currently that of a superior rebounder and less porous defender. But Tennessee’s still a worse defensive team with him on the court – far worse at the rim – and he doesn’t make up for it as much offensively.

The story is not as simple as this, though. Evan Miyakawa’s Bayesian Performance Rating (BPR) has Huntley-Hatfield as worth about five points more per 100 possessions. Traditional Box Plus-Minus gives Huntley-Hatfield a 3.6 points per 100 edge. Even Torvik’s Box Plus-Minus, an updated version of the original, pushes Huntley-Hatfield out in front by 3.1 points per 100. By any measure you look at, Huntley-Hatfield should be getting more of Plavsic’s minutes.

But I still feel bad. I still feel like I’ve been far more rude to Plavsic online than any sort of writer decorum calls for. So: I apologize, Uros Plavsic, for the rudeness. You do have some amount of value, and it is not your fault specifically that the lineups they include you in drive me nuts. Tennessee is almost exactly as good with one of these two players on the court (+24.8 per 100) as they are without either (+25.1). It doesn’t matter as much as I claim it does.

15. Recency bias.

The tendency to weigh the latest information more heavily than older data.

  • Against Colorado and Texas Tech: 12-for-64 on threes (18.8%)
  • Against the other six opponents: 61-for-160 (38.1%)

Here’s the thing: literally no team ever is going to shoot 19% from three for an entire season. Tennessee entered this season with four players who have posted a season of 38% or better from three and added Kennedy Chandler and Zakai Zeigler to that roster. By pretty much any measure you choose, Tennessee has the most shooting it has had since the 2006-08 Bruce Pearl teams. The 2006-07 team, which had Chris Freaking Lofton and eight total players with 18 or more made threes, posted seven games with a 25% hit rate or worse from three.

It is a long season. College basketball is a goofy sport where there are no rules about what basketball you use at which venue. Don’t give up yet.

16. Salience.

Our tendency to focus on the most easily recognizable features of a person or concept.

The example provided in the Bias Chart is being worried about dying at the paws of a furious lion ripping your torso to shreds rather than the much more statistically likely thing, some doofus missing a red light and turning your car into a museum exhibition. You could run this one any number of ways, but what if we have zoomed in too far on Olivier Nkamhoua’s obvious, goofball mistakes while failing to notice he has taken a legitimate step forward?

Olivier Nkamhoua, 2021-22:

  • Third on the team in scoring
  • Second on the team in blocked shots, #9 overall in the SEC
  • Team leader in rebounds, #8 overall in the SEC in total rebounding percentage
  • Second-lowest Defensive Rating among rotation members
  • Team leader in 3PT%, minimum 10 attempts
  • Team leader in midrange FG%, minimum 10 attempts
  • Fifth on the team in PORPAGATU (aka Value Over Replacement)

You get the point. This is one of Tennessee’s best players and a deserving starter for now; if Josiah-Jordan James ever gets things fixed, I’d be open to Nkamhoua being a sixth man to relieve the rotational backlog, but I would prefer to not hear something as absurd as “I’d rather see Plavsic” in my mentions going forward.

17. Selective perception.

Allowing our expectations to influence how we perceive the world.

(looking around at America) Can’t figure out how this one fits in sorry

18. Stereotyping.

Expecting a group or person to have certain qualities without having real information about this person.

In the season preview, I said this about Zakai Zeigler:

“The problem with Zeigler is going to be an obvious one when you hear his height and weight: 5’9”, 167. Zeigler was a late add to this year’s team, so he didn’t even get a summer in the weight room; I would be very worried about how well he can hold up in ball-screen defense as well as isolation situations. That’s going to put a hard cap on his playing time potential against any non-Quadrant 4 opponent.”

What is the wrongest you have ever been, and were you this wrong?

I have enjoyed the Zakai Zeigler Show so much that I advocated for a crowdfunded statue midway through the first half of Act of God and that somehow doesn’t feel ridiculous to imagine. Zeigler is simply the most exciting non-star (for now) Tennessee has maybe ever rostered, a player that is going NASCAR speeds from 40:00 to 0:00 and never, ever stops. At one point during Tuesday’s game, Zeigler was matched up with a player eleven inches taller with him and kept that player from scoring. I love this player and want him here forever.

19. Survivorship bias.

An error that comes from focusing only on surviving examples, causing us to misjudge a situation.

A Barnes-era example: having a complete outlier of a midrange FG% (nearly 47%) in 2018-19 and believing it validates taking 15 mid-range jumpers a game rather than correctly seeing it as a huge outlier. (It appears this has mostly been fixed, though, so no need to harp.)

A this-week example: if Santiago Vescovi hits that wide-open three to give Tennessee a 47-44 lead, do we start talking about how tough or full-of-heart or whatever crap Tennessee basketball is because they went 5-for-32 from three in regulation instead of 4-for-32? Or do we correctly call this the Act of God game and move on with our lives?

Finally,

20. Zero-risk bias.

Sociologists have found that we love certainty – even if it’s counterproductive.

This one closes with “eliminating risks entirely means there is no chance of harm being caused.” It is fundamentally antithetical to the sport of basketball, a game where your goal is to only miss half your shot attempts. Even the very best college basketball teams will miss six out of ten threes. You have to get used to a lot of clanks. That’s the risk. The upside is immense when it works; when it doesn’t, you do feel a bit stupid.

I would advise that Tennessee not slide back into the warm certainty of mid-range jumpers. This team has about 2.5 players that should be allowed to take them; everyone else, rim-and-threes. Am I committing myself to my own brand of zero-risk bias by claiming this is superior? Of course. That’s how we get to do this entire chart again.


The point of this post is that Tennessee played two basketball games and played utterly terrific defense in both. They’ve played eight games of basketball and have held the opponent below 50% at the rim in seven of them. That is a fabulous rate that every team in America would kill for. Yet two outlier shooting games have killed the offense’s own narrative. You can’t expect anyone to shoot 20% or worse every time out against good opponents, just like you can’t expect anyone to shoot 50% or better every time out against the bad ones.

It is what it is. Saturday begins a new week and a new opportunity. Might as well re-invest; after all, many have called this blog Stats by Ostrich Effect for a reason.

If you like the format of these or want to see changes, let me know at statsbywill@gmail.com.

One thought on “The Bias Chart

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s