Do 3-Point Shooting Teams Win in the Tournament?

The Setup

I know we usually focus on the NCAA, but stick with me for a second.  Did you know that Steph Curry has the record for taking and making the most 3’s in a career?  He currently has 3640 made 3’s, which is nearly 700 more than second place Ray Allen, or James Harden or Reggie Miller.  Two of those guys played full careers and 350 more games than Steph.  In fact, five of the top ten 3pt made shooters are still active.  That’s nuts.  That means they are jacking up three’s all over the place right now at a rate that has never been done before.  And that idea has bled over into the college game.

Here’s why.  There is a pretty basic analysis that goes something like this:  You only have so many shots to take in a game (about 60), so if a team shoots all high percentage (50%-ish) two-point shots then they will score around 60 points.  But if they replace those 2pt shots with lower percentage (35%-ish) 3pt shots, they are compensated by getting 50% more points on each make.  Meaning that in aggregate you will score more points.  In this example – 63 points. 

Alright, for the Army guys in the room I’ll slow it down.  More 3pt attempts means more total points and that means more wins.

And if a team focuses on volume shooting three’s it may even become better because you will get more comfortable and potentially shoot a higher percentage.  For example, Steph shoots 2% better than anyone else ever. 

The Question

So, I get that this works in large numbers, for instance over the course of a season but I wonder how effective this is in college basketball?  Particularly does high volume 3pt shooting translate to deeper runs in the NCAA Tournament?

Part 1 – High 3pt Percentage

In order to examine this idea, we have to first get a baseline.  Different seeds tend to go further in the Tournament, so comparing a 1 to a 16 doesn’t make sense as in the last 20 years a 1 has averaged 3.2 wins and a 16 averages only .02.  In fact, we aren’t even considering 16’s because only 2 have ever won a non-play-in game.  We broke the rest into tiers: 1, 2-4, 5-7, 8-12, & 13-15.  Teams in those groups average roughly the same number of expected wins – same ballpark, at least.

Then we separated the teams into 5 quintiles by 3pt shooting, the top 20%, 60-80%, 40-60%, etc.  And here’s what we found:

  • For the 1 seeds, being in the top 40% of 3pt shooters made a slightly above average difference.  They won 0.2 games more than the mean. 
  • Seeds 2-4 and 5-7 do not indicate any notable advantage by being a good 3pt shooting team. 
  • Seeds 8-12 are marginally better if they shoot the three well
  • And the last group, 8-12 actually tend to perform worse in the Tournament if they have relied on a high 3pt shooting percentage all season.

Part 1 Conclusion – Don’t Fall in Love with Good 3pt Shooting Teams

We were really hoping to find a smoking gun here.  I mean, if you shoot a higher percentage in your 3pt shots, it would seem to be reasonable that you would be able to win more games.  But so far the data is mixed as best.  By the way, this highlights the fact that trying to find a single metric to serve as a predictor of tournament performance is hard to do.

Claiming, “I like this team to make some noise; they’re one of the best 3pt shooting teams in the country” is not good reasoning when filling out your bracket, especially for the lower seeds. 

At the NCAA level and particularly in the pressure cooker of the Tournament, a great 3pt shooting percentage is not the indicator that is going to help you win your bracket.  So let’s dig deeper.

Part 2 – The Deuce

“Live by the three; Die by the three!”

Ok, so what about volume shooting of the three.  Let’s isolate that factor and see if you jack it up from downtown more often, will you advance in the Tournament farther than expected. 

The short answer is volume shooting is not really significant for any quintile.  But there are a couple of interesting notes.

  • Only 8 of the last 80 1-seeds were in the top 20% of volume 3pt shooters.  And 4 of those 8 were Villanova.  As a group, they were about on par for expected wins. 
  • Seeds 5-7 are the only group with a notably above average win rate with the highest volume 3pt shooters.  Even then they only win 0.15 more games than expected.
  • 8-12 Seeds: 65 teams shot in the 60-80% quintile, not the highest volume group but above average, and 3 of these teams reached the Final Four.  That is a rate 87% higher than expected. 
  • For 8-9 seeds, you can expect 9% go to the Sweet 16, but if you volume shoot in the 60-80% quintile your rate jumps to 23%.  That’s pretty cool and now we’re getting somewhere.

So far, it doesn’t look good for the Splash Brothers school of thought.  You can’t say, “My boys shoot a ton of threes and they are going to get on a roll” or “We shoot such a high percentage with the three that we are definitely going far in the Tournament.”  In a one-off situation, that may occur, but it’s not likely to result in more wins than average in the NCAA Tournament, based on historical data. 

But dang it, I can’t leave this idea alone.  There has to be something to this three-point shooting trend that will help us predict the perfect bracket.  And there is.  But I’m going to need a day or two to write Part II of this article up.  Until then, wait anxiously.  Hit refresh often.  I’ve missed you guys.

2 responses to “Do 3-Point Shooting Teams Win in the Tournament?”

  1. Hugh Stephenson Avatar
    Hugh Stephenson

    Regardng my march madness pool drafting players …
    I have a running excel spreadsheat that shows the avaerage number of games played in the NCAA tournement (s 1990)regardless of seeding.

    ST PETER’S-4
    UNC 3.74
    KANSAS 3.43
    MICHIGAN 3.40
    DUKE 3.32
    KENTUCKY 3.25
    UCLA 3.20
    FLORIDA 3.18
    HOUSTON 3.0
    LA SALLE 3.0
    SYRACUSE 2.94
    Gonzaga 2.86
    BAYLOR 2.80
    MIAMI FL 2.75
    OREGON 2.75
    BUTLER 2.69
    ARIZONA 2.68
    AUBURN 2.67

    268 total teams

  2. Hugh Stephenson Avatar
    Hugh Stephenson

    Good Morning. Excellent analysis as always.
    For the past 25 years I have been involced in a really cool March Madness pool:
    12 guys each put 100$ in the pot.
    We each draft 15 players partcipating in the tournenment -not play in games.
    The draft is held on WED EVENING prior to the thursday opening round.
    We draw numbers out of a hat for draft order.
    The first 5 rounds you get 1 minute to pick a player.
    Rounds 6-10 you get 3 minutes to pick a player.
    Rounds 11-15 you get 5 minutes to pick a player.
    The person who picks 12th gets 2 picks (the end of round 1 and beginning round 2) and we work our way back down the order to the guy who picked first. Rinse and repeat proces.
    If after all picks are in someone realizes a player has been injured or arrested the person who realizes this gets to replace that player with anyone who was not drafted.

    Its an interesting way to straegize your luck especially when you are last inthe draft order.
    Some guys load up on the chalk teams (seads 1-4) while other players pick top scores of teams with lower seads. It all comes down to how many games you think a particular team will play in the tournement and not actually picking a perfect bracket.

    So with that being said I am always looking for what regular season stats have predicted how far a team might go, ie playing more games in the tourney