Millie Prior and the Math Behind Davidson's Defense

It's well known in analytical circles now that even the very best free throw shooters -- from Stephen Curry to Paige Bueckers to Foster Loyer -- can only hit up to about 95% of their attempts over a season. Ridiculous streaks are, of course, possible. Currently, Bueckers has missed exactly ONE free throw attempt since Christmas. That's good for a 98% hit rate.

At some point, though, the law of large numbers kicks in. (You might know it as the "Law of Averages") TLDR: the larger a random sample is, the more reflective the results become of reality. There's a human limit to how many shots one can hit without missing. In fact, the very game before Bueckers's current ridiculous streak, she was 2-4 from the stripe against USC. Her Huskies lost by 2 on a dramatic miss at the buzzer. Maybe she simply decided that missing shots was for bums after that, but I'd be willing to bet it was pure randomness, or Dumb Luck, or Fate. You can call it whatever you like.

Likewise, there's a human limit to how poorly you can make your opponent hit jump shots. Even the very best defensive schemes can only limit their opponents to, say, 25% from deep over the course of a season. Why? Well, defenses generally have far less control over whether their opponent hits a three than a two. We call centers Rim Protectors for a reason; the best ones influence every shot at the rim. The result of this is that some teams ride stretches of their opponents' ice cold shooting -- i.e. a random sample that comes out poorly -- into making you think they're a lot better than they actually are.

The most dramatic example I can think of across A10 MBB and WBB are the Saint Louis Billikens men's side. In non-conference play, SLU's opponents shot a blistering 36.2% from three. We were sure of 2 things: their defense sucked, and they weren't yet a contender.

Fast forward to conference play. The evening of January 29th, to be precise. SLU beat VCU in pretty convincing fashion at Chaifetz to get to 6-2 in the A10. Naturally, then, our collective reaction was that the Billikens had figured it out and were heading for a Double Bye and title contention. However, SLU's opponent three point percentage in conference -- just 21.3% at that time! -- led to me sending this Bluesky Tweet (Skeet?). Since then, SLU has lost 4 out of 5, and their opponents have again set the nets on fire with a 39.6% hit rate from distance. They're 7-6 in conference, and a double bye feels like a long shot. My bad.

I say all of this because if you look at the season stats for this Davidson team, one thing jumps out immediately: Davidson is the very best 3PT defense in the entire country, at least as of the evening of February 21st, 2025. Their opponents are shooting only 24.8% from deep, almost 7% below the national average.

After everything I've just said, you might think that Davidson is set to ride the regression rollercoaster back to reality. However, I have a good reason they've avoided it -- and the answer is not Magic Beans that give you favorable variance. It's a little more interesting than that, and it involves the numbers at the bottom of that "DEFENSE" list.

(Mostly) Basic Math

Pop quiz: how many of Steph Curry's 3PA this season have been assisted?

The answer is, of course, 68.7% of them. The point is that even the absolute best 3PT shooters make more threes off of assists than not. Even in 2015-16, the Steph Curry Season to Rule Them All -- and maybe the single best offensive season in NBA history -- 53.9% of Steph's makes from distance were assisted.

Ever since the Kid from Davidson dragged it kicking and screaming into reality, the entire basketball world has realized that 3 is undoubtedly a larger number than 2. Three point attempt rates have obviously been rising for years, but Steph became the somewhat unwitting face of all that. The NBA's 3PAr is now so high he routinely faces charges of "ruining the league" -- even though all of those people surely marveled at him every night in those halcyon days.

So how do you design the perfect defense for the modern era of basketball? To me, you have to account for several things we've discussed that are axiomatically true:

Number one: If you want to take as many threes as possible, then you want your opponent to take as few threes as possible.

For one, good shooting teams usually get more expected value (EV) out of a 3PT shot than a 2PT shot. Consider 2022 Davidson MBB. They were extremely good at shooting both twos (34th nationally) and threes (4th). Therefore, their EV for a 2PA = .542*2 =1.08. Their EV for a 3PA = .385*3 =1.16.

Using the DI averages for WBB in the 2024-25 season, the EV of a 2PT FGA is 2*.457 = 0.914. For a three, it's 3*.313 = 0.939. Simple math. The better an offense's shooting is, the wider this gap (usually) gets. Obviously, there are exceptions based on style, but with the way offense is currently played, that gap is worth 2.5 points per 100 possessions.

And, as we've said, you have more control over whether a 2PT attempt goes in than a 3PT attempt. 3P% is a much noisier stat over a small sample, such as one basketball game. That's the other reason you want to erase them from your opponents' arsenal as much as possible: it becomes harder for them to ride a shooting heater to a win.

Secondly, it's better when your opponent takes shots off the dribble than off the catch. That EV above was for all three point jumpers. I don't have the data for assisted three pointers, but you can bet it's probably much higher. An assisted three for a good shooting team is the best shot in basketball, other than one assisted at the rim, and...we'll get to that.

Think of it in this way: at some point in 2021-22, opponents probably reckoned that Luka Brajkovic one-on-ones in the post were better to allow than open threes off the catch for Lee, Loyer, or Jones. You can't blame them! However, Luka then decided to hit SIXTY FOUR percent (64%) of his twos that season, which made everything a wash in the end. The EV of a Luka 2PT FGA was 1.28, which is insane. Once he decided to hit 40% of his threes, it was Over. (He's still underrated, but that's another article.) This is an extreme example, but just know you have to be ridiculously efficient for unassisted shots to be better than assisted ones.

How do we accomplish those two things? How do we get a "No Threes + No Assists" defense?

1. Never help off shooters, so they won't get open shots. Never. The other team can't have a drive and kick game, because helping one pass away dramatically increases the odds of an open shot, which is a lot more efficient. More players can hit a wide open three than beat a good defender in isolation.

2. Switch every screen. All of 'em. That way, pick and pop jumpers aren't open and pick and roll buckets at the rim aren't open. We want the PnR ball-handler to shoot, not the roller or someone in the corner who was helped off. Stay connected to screeners and switch off the ball too, so shooters can't get open off pin downs.

3. Force as many midrange jumpers as possible. Rim and three attempts are more efficient and more often assisted than long twos. Our opponents' shot quality needs to be as bad as possible because, once again, the number of great 3PT shooters is now much higher than great back-to-basket or midrange shooters.

   
   

That -- exaggerated greatly above -- is Davidson's defensive philosophy. (Clearly there are exceptions -- you often need to stunt and recover or try to force more turnovers, etc.)

The Cats have the lowest opponent assist rate in the country. Only 39.2% of their opponents' makes are assisted. Tied closely to that is the fact that only 25.7% of their opponents' FGA come from deep, which is 6th lowest in the nation.

That, in essence, is how Davidson grades out as the best 3PT defense in the nation. They decided only jokers and bums allow other teams to take uncontested jump shots, and they eliminated any tendency that would cause them to surrender such a look.

Scheme is one thing. It's nothing without personnel, and Davidson's scheme asks its players to do a ton of work on defense to make it all come together. Your guards need to be tall and strong. In addition to having to guard the perimeter, they're going to be left on islands in the post in isolation. If you were wondering why Davidson's shortest starter is 5'10" -- and why there's only one player below 5'9" on the entire roster -- there's your answer. A big posting up a guard or wing is still better than that big passing out to an open shooter after getting doubled, and that's not an insane position to take because Davidson's guards and wings can all switch and hold their own.

Charlise Dunn and Candice Lienafa are excellent examples of how athletic players become weapons in this scheme. Dunn's defense is particularly incredible for how unheralded it is; she can guard essentially every position and has the wingspan to erase passing lanes while staying attached. She's averaging 1.5 steals and 1.0 block a game. Lienafa is likewise a weapon for many of the same reasons. She plays much bigger than her listed 5'11" in the post (just watch her tape against Zahirah Walton and George Mason), but has the quickness to stay in front of guards as well.

There's one more way Davidson's length helps: on the glass. Rebounding well doesn't make you a great defense, but it gets you pretty far. Also consider that missed 3PTs produce longer rebounds, which means more offensive rebounds. Giving up fewer 3PT FGA thus makes for better defensive rebounding.

The linchpin of any great defense, though, is the center. This one is no different. That's why it's time to talk about Millie Prior and what makes her so important.

Millie Prior

We've at least rejected the idea that teams should play Brexit Basketball, i.e. pack it in and simply force teams to shoot over them. That's too boring. You can influence your opponents' 3PT shooting plenty and significantly reduce 3PT variance in the process.

If you're going to take away the drive and kick game completely, though, you better have someone who can protect the rim at a high level as the last line of defense. The key to this philosophy is that 2PT FGA are easier for you to influence than three point jumpers -- you're generally always closer to the player taking the shot. That's true if you have a great center to take advantage of this fact.

It takes a special kind of center to do everything without help. A center with high basketball IQ. A center who can hold their own against perimeter players. A center with great reach who also can't get backed down in the post.

That's Millie Prior.

Stats don't quite capture her entire impact. But as of writing, Millie Prior has the following ranks across the A10:

• 1st in Blocks Per Game at 1.4

• 2nd in Block % at 6.3%

• 3rd in Defensive Box Plus/Minus at 4.2

It's very difficult for post players to move Prior. You can't back her down or out-tough her. However, it's most important for the scheme that Millie's length and IQ allow her to impact shots from far away. If a guard does get switched onto her, she can get deep, stay in front, and make them shoot over her -- which is generally always a bad shot. The next option is driving right at her, which also usually doesn't go well. She's averaged nearly two blocks a game the last two years for a reason.

The definitive Millie Prior game for me was actually her first game back from a shoulder injury. It's also one Davidson lost -- albeit in heartbreaking fashion. It was when Davidson pushed NC State to the brink, falling 59-57.

The previous outing against Ball State, the Cats went on the road to face a high octane offense and likely tournament team. Without Millie, the Cardinals put up 27 three point attempts -- taking over half of their shots from distance. That's far above their season average of 40%. They pushed the tempo well above Davidson's preferred slow pace. The next game would become the turning point.

In her return, Prior had 9 rebounds and 5 blocks. NCSU attempted only 11 threes. More impressively, the Pack, who've hit 51.2% of their twos this season, went 18-52 from inside the arc. From the midrange, there were 7-28. Putrid. At the rim, they were 11-24, which is not horrific but pretty bad for a great offense.

Over the last month, Davidson has been completely locked in. With Prior and a healthy Tomisin Adenupe providing great minutes off the bench at center, Davidson's defense hasn't allowed more than 51 points in February. Opponents are shooting an absolutely ghastly 31.5% on twos over the Cats' last 5 games. Think about that. On twos! Lest you think it's a matter of weak competition, consider that George Mason went 15-43 on twos Thursday.

And while Davidson routinely puts up 25 or so threes a game -- hitting them at a 36.2% rate in conference play -- you'll be lucky to even put up 15. Davidson hasn't allowed an opponent to hit more than 6 threes in a game in conference play, and that was against Richmond, which is quite literally the best shooting team in America. They only attemped 13 threes. In fact, every single one of Davidson's opponents has attempted 16 threes or less with the exception of Rhode Island. The Rams attempted 24, but that was quite intentional and an instance of great game planning from the coaching staff.

---

If this defensive scheme sounds difficult to pull off -- you're right. It takes a carefully constructed roster.

The next time you see the number one team in men's basketball playing, though, pay careful attention to how they guard the perimeter. It might remind you a little of Davidson.

Auburn's shortest starter is 6'4". They, too, have a veteran center who's an incredible defender in Broome. And they, too, do not allow assisted field goals. They have the 3rd lowest opponent assist rate at 38.8% and 37th lowest opponent 3PAr at 34.3%. The result is that their opponents are hitting just 29.3% of their threes, which is 11th best in MBB. They also block a ton of shots. They're committed to eliminating the drive and kick, which is part of the reason Alabama's spread struggled so much against them last weekend. Auburn's made more headlines for record breaking offensive efficiency, but their defense certainly raises their floor.

Like any scheme, there are weaknesses. Yes, shifty guards who can blow by you. Yes, guarding aggressively one-on-one leads to a lot of fouling (don't check Davidson's FTr on defense). And yes, transcendent bigs like Luka Brajkovic can bully you a little.

However, I'll end with food for thought: it's much harder to break down a defense when you have to have a certain type of player to do so. Players like the ones above don't grow on trees the way 3 and D shooters do nowadays. Not everyone can do those things, and that's raised Davidson's floor to new heights.

We'll see just how far mathball can get you, but it's been a pretty fun ride so far.