2024 Bracket Picks Year in Review
We review the performance of our 2024 NCAA bracket picks, including what went right and wrong with our data-driven strategies and analysis.
The 2024 NCAA tournament saw No. 1 Connecticut become the first repeat champion since Florida in 2006 and 2007. This time, Connecticut started off as the No. 1 overall seed and went wire-to-wire, setting the record for largest average margin of victory (23.3 points per game) for a champion.
Outside of Connecticut’s second dominant tourney run in a row, however, the 2024 March Madness campaign played out differently than last year:
- 2023 was relatively upset heavy tournament, and a No. 4 seed was the top seed to reach the Final Four.
- 2024 featured a title game between two No. 1 seeds, and only two teams worse than a No. 5 seed advanced past the Sweet 16.
Also for the second year in a row, our customized brackets for PoolGenius subscribers were heavier on Connecticut than the public was. Our pre-tournament analysis gave UConn the best odds to win the tournament by a solid margin, while several other top computer ratings systems were highest on No. 1 Houston this year.
Connecticut’s relatively high popularity as an NCAA champion pick in bracket pools meant that our Best Brackets with UConn as champion were targeted mostly for smaller pool sizes. Still, that strategy, combined with our high opinion of Connecticut in general, was enough for our 2024 brackets to significantly outperform the public overall.
Early subscriber survey responses are reinforcing that outcome, and currently have us on track for one of our higher subscriber prize win rates in recent history.
Reminder: If you used the NCAA Bracket Picks product this year, please complete our subscriber survey if one is available to you, so we can collect real-world data on how our brackets performed.
PoolGenius 2024 Bracket Performance Summary
Our pre-tournament data had Connecticut in a tier of their own as the favorite, with win odds nearly double those of other No. 1 seeds Houston and Purdue. They were, though, also the most popular public pick to win it all.
Here’s how that dynamic generally impacted our optimized brackets in 2024:
- In pools that use the traditional scoring system (1-2-4-8-16-32), our Best Brackets at small pool sizes favored No. 1 Connecticut as champion, with No. 1 Purdue being the runner-up at smaller sizes, on account of Houston being the more popular public pick to advance that far relative to Purdue.
- As pool size increased, our Best Brackets went more contrarian, with some champion and runner-up picks that included No. 2 Arizona, No. 2 Marquette, and No. 4 Auburn. Connecticut still showed up on some alternate brackets for mid-sized pools, but was less often paired with Purdue as the runner-up. This is in contrast to last year, where Connecticut was a value champion that was showing up as pool sizes increased.
- Brackets for other types of pools with non-standard scoring varied in how often Connecticut was picked as champion. UConn showed up at least as often in other round scoring formats, and in scoring that included both Round and Seed scoring, but other champions were more likely on some brackets for purely seed-based scoring (i.e. without round-based scoring).
- Our Best Brackets at smaller pool sizes (e.g. 35 entries and under winner-take-all pools, and in slightly larger pools where finishing in the top 3% was solidly in the money) finished at the 98.8th percentile this year in ESPN’s Bracket Challenge, very nearly a top 1% outcome with those brackets.
April 24th Performance Update with Survey Results
With several weeks of survey results now calculated, here is how the numbers have shaped up for 2024.
Overall, subscribers won a prize in each individual pool 36% of the time, versus an expectation of 11% based on pool size and payout information. That 3.2x is in line with, but slightly higher, than our average yearly Bracket Picks performance.
As expected, that win rate was dependent on pool size, with pool sizes of 100 entries or fewer showing the highest win rates in 2024. For example, in pools with between 31 and 50 entries, 51% of our subscribers reported winning a prize, compared to an average expectation of 11%.
At between 101 and 1,000 entries, our subscribers still won at a higher rate than the public expectation, though the edge was not as high (around 1.5x the public average). At pool sizes above 1,000 entries, the win rates were below the expected average, which is not surprising given the dynamics that we detailed below.
The bracket performances were also relatively higher in standard scoring rules pools (winning 3.6x more often) and other pools with round-based scoring only (winning 3.3x more often). Meanwhile, there was no edge in 2024 in seed-based scoring formats, where fewer upsets and the lack of notable key underdogs going on runs contributed to lower scores.
Overview of the 2024 NCAA Tournament
- No. 1 Connecticut (East) followed up their dominant run last year with an even more dominant one in 2024, winning the tournament as the top overall seed, with no game decided by fewer than 14 points.
- No. 11 NC State was the Cinderella darling of this tournament, following up a miracle ACC tournament run to get into the NCAA field. After a last-second victory in the play-in round, NC State reeled off a four-game run through the South region, where they beat Texas Tech, Oakland, Marquette, and Duke (for the second time in a few weeks). It was the first Final Four for NC State since the miracle run to the title in 1983.
- No. 1 Purdue exorcised the demons from last year’s First Round loss to Fairleigh Dickinson, by going all the way to the NCAA title game. It was also Purdue’s first Final Four appearance since 1980.
- No. 4 Alabama got a thrilling win over No. 1 North Carolina, and followed it up by beating No. 6 Clemson in a matchup that seemed more fitting for a College Football Playoff.
- No. 14 Oakland was the biggest early round Cinderella, knocking off Kentucky with a lights-out shooting display from Jack Gohlke.
What a Winning Bracket Looked Like in 2024
Based on how this year’s NCAA Tournament played out, our subscribers likely saw scoring totals in their pools that were higher, on average, than in most years, especially with two No. 1 seeds making the Final Four.
Based on public picking trends, the typical bracket had between one and two Final Four teams correct in 2024, most often Connecticut and Purdue.
To get an idea of what kind of brackets won pools this year, we looked at the top 10 scoring entries in a 300-entry pool with standard 1-2-4-8-16-32 scoring. Here’s what we observed:
- All of the top 10 brackets had No. 1 Connecticut beating No. 1 Purdue in the title game.
- None had No. 4 Alabama or No. 11 NC State in the Final Four—and this is a fairly large pool!
- Six had Alabama beating North Carolina to reach the Elite Eight.
- Two had NC State making a surprise run to the Elite Eight, but none had both Alabama and NC State in the Elite Eight together.
- Seven had NC State losing in the First Round, and one had both Alabama and NC State losing in the First Round (and still finished in the top 4%).
- These top-scoring brackets picked 22.2 games correctly in the First Round, and 12.4 picks correctly in the second round. That’s better than the public average, and similar to our Best Brackets average in those two early rounds (21.8 and 12.9). It also shows how the winners mostly had top contenders advancing to the Sweet 16 and beyond.
In some small pools smaller than 300 entries this year, it probably became even less important to have either Alabama or NC State in the Elite Eight, or alternatively, to have Purdue in the title game vs. just the Final Four.
PoolGenius 2024 Bracket Performance Detail
Here is how our bracket picks compared to the public for the 2024 tournament in each round:
| Bracket Type | Correct R1 Picks | Correct R2 Picks | Correct Sweet 16 Picks | Correct Elite 8 Picks | Correct Finalist Picks | Correct Champ Picks |
|---|---|---|---|---|---|---|
| PG Best Brackets For 1-2-4-8-16-32 Scoring | 21.8 | 12.9 | 2.97 | 1.86 | 1.30 | 0.55 |
| PG Best Brackets For All Scoring Rules | 21.2 | 12.0 | 2.89 | 1.76 | 1.25 | 0.58 |
| All PG Brackets For 1-2-4-8-16-32 Scoring | 21.7 | 12.5 | 3.04 | 1.35 | 0.97 | 0.38 |
| All PG Brackets | 21.4 | 11.7 | 2.92 | 1.29 | 0.94 | 0.39 |
| The General Public | 21.2 | 10.3 | 2.71 | 1.10 | 0.69 | 0.30 |
As it turns out, not only did our Best Brackets for standard scoring outperform the public in every single round, but the average of all of our brackets (including alternate/portfolio suggestions) also finished ahead of the public in every round, with the exception of our Best Brackets For All Scoring Rules only tying the public in the First Round.
Let’s take a deeper look at what went right and wrong in 2024.
What Worked and Didn’t Work in Our 2024 Brackets
Generally, a successful bracket picking strategy (especially for larger bracket pools) requires making some calculated bets against the public.
The logic that ultimately determines which teams we “bet on” in our customized brackets for subscribers is driven in part by each team’s tournament rating, which often includes an adjustment we make to its full-season power rating based on factors like lineup changes and injuries.
In retrospect, we had some hits and misses with our ratings adjustments this year, as will happen every year, but our conviction in Connecticut was key.
(As a quick reminder, we provided subscribers with all our adjusted tournament predictive ratings in our NCAA Bracket Picks product. We also provided a detailed writeup containing our thoughts on the 2024 bracket and associated pick strategy.)
Because our algorithms customize bracket picks based on a pool’s size and scoring system (e.g. upset bonuses will change recommended picks to include more upset picks), the optimized brackets our subscribers generate can differ quite a bit from pool to pool. Nevertheless, there are some common themes that likely impacted 2024 results for the majority of our subscribers.
1) Connecticut Dominates the Tournament, and Our Winners (Again)
Overall, we had a much higher percentage of brackets picking the NCAA champion correctly than the public did:
- 55% of PoolGenius Best Brackets for traditional scoring had Connecticut as champ
- 39% of all PoolGenius customized brackets had Connecticut as champ
- 30% of public brackets had Connecticut as champ
It’s easy to say in retrospect that it was obvious Connecticut was the best team. But while they were the betting market favorite at the start of the tournament, there were plenty of pundits and rating systems out there that had Houston or Purdue rated higher.
In fact, in our Tuesday update comparing other rating systems to our adjusted ratings and odds, we noted that only looking at ratings on sites like Ken Pomeroy and ESPN’s BPI might lead one to conclude that Connecticut was a terrible champion pick to make. (For example, ESPN BPI gave Houston almost three times the chance to win the tournament as Connecticut, 35.3% to 12.8%.)
Indulge us while we re-print in full what we wrote about our position on Connecticut on Tuesday before the 2024 First Round began. Because while our methods are by no means perfect, this is a strong example of how the deeper level of analysis that we do can deliver an edge.
The Case For Connecticut In 2024
So why do we have Connecticut as our tournament favorite, even putting their odds above the betting market’s implied odds, while all the other systems favor Houston? After all, our unadjusted predictive ratings heading into the tournament also have Houston No. 1, and Connecticut in a virtual tie with Purdue and Arizona for No. 2.
If we ran bracket simulations without accounting for anything else, we too would probably have Houston with a far-above-market win projection, and Connecticut below.
But we ended up adjusting Connecticut’s rating up, and Houston’s down.
Here’s the biggest reason why we adjusted Connecticut up, and you can decide if you agree with our logic. Regular season games missed by at least one starter:
- Houston: 1
- Purdue: 0
- Arizona: 0
- Connecticut: 11
And these weren’t just any absences. Most either involved Stephon Castle or Donovan Clingan, two of Connecticut’s three highest usage players on offense (along with Tristen Newton), and their dominant defensive presence on the interior in Clingan.
Before we get into more numbers, just knowing that Connecticut is rated similarly to these three other teams (who collectively had only one game missed by a starter) in our unadjusted power ratings, while they played without their top lineup for over one-third of their games this season, should illustrate why the market (and us) have Connecticut as the tournament favorite and not Houston.
Here are the splits for Connecticut with and without all five starters:
- With all five starters: 22 games, +26.2 power rating
- Without a starter: 12 games, +17.7 power rating
You can see in our team notes that as of today (Tuesday), we’ve settled on a +24.8 rating for Connecticut, baking in some regression from their performance when healthy. But we are willing to take a stand that this Connecticut team is a clear step above what its end-of-season, unadjusted power rating implies once you dig in to the details.
Even the betting market may be a bit lower than it should be if some influential market participants are primarily tethered to unadjusted full-season numbers, including games in which Connecticut played relatively worse (but still at roughly a No. 3 seed level) when missing a key player.
As it turns out, Connecticut’s average Game Score in the NCAA Tournament was +36.2, a staggering number even larger than the regular season rating at full health.
If anything, we are kicking ourselves for not taking a bigger stand against the betting market on Connecticut, considering that at full strength they had posted a Game Score of +27.0 or better in nine of the their final 11 games entering the tournament. While we respect the market, this was a case where our read on a team was more informed than the market average.
Still, the position we did take is the biggest reason for the high 2024 bracket pool win rates we’ve seen reported so far in early survey responses.
2) Purdue As Runner-Up Was Also Big
Of course, the Connecticut call was important, but it’s likely true that a high percentage of subscriber brackets that cashed this year not only had Connecticut as champion, but also paired them with Purdue as the runner-up. While getting the champion pick right is the most important part of a bracket (using standard scoring), if the champion ends up being a reasonably popular No. 1 seed, hitting the runner-up is also key.
At smaller pool sizes, which make up the majority of the pools our subscribers play in, Connecticut being paired with Purdue ended up being important.
Our position on Purdue was driven partly by us being lower on Houston due to some adjustments. Although our advancement odds were similar for these two No. 1 seeds on the right side of the bracket, two factors came into play:
- Houston was the more popular runner-up pick in bracket pools, so there was contrarian value in avoiding them (all else being equal).
- Houston’s path was a bit tougher early on, as we saw both Texas A&M and Duke as tougher challenges than the No. 8/9 winner facing Purdue, which was also in a region with an injury-impacted No. 4 Kansas.
While we could not have foreseen Jamaal Shead’s injury, what we did downgrade Houston for was some injury uncertainty and depth concerns due to injuries to several rotation players. Unlike Connecticut, Houston was likely not as good as their regular season numbers with the players they had entering the tournament. Those concerns proved valid based on Houston’s inability to close out Texas A&M and their loss to Duke.
Combinations Of No. 1 Seeds Aren’t As “Chalky” As You May Think
We will note that while it seems far too obvious and chalky to many people to pick two No. 1 seeds as your title game matchup in a bracket, you actually would expect the percent of brackets picking Connecticut vs. Purdue to be fairly low.
There are different ways to estimate that percentage, but one quick and simple method is to multiply the 30.4% of the public picking Connecticut as champ times the 12.2% picking Purdue as runner-up. That gives us an estimate of only 3.7% of brackets in total.
The true number could be higher or lower than that, depending on how the public’s biases affect the way the two picks are correlated across brackets.
3) Alabama Saves the West from North Carolina
While none of our Best Brackets had No. 4 Alabama in the Final Four, the Crimson Tide beating North Carolina was a fairly important factor in our overall performance this year.
We noted in our Rooting Guide that North Carolina was our biggest fade of 2024, in terms of being a team that was a popular pick to win the title, but did not appear as even the runner up in any of our Best Brackets or alternate brackets for standard scoring.
After Arizona (our most common Final Four pick out of the West) was upset by Clemson in the Sweet 16, the path appeared to open for North Carolina, and we started to get that uh-oh feeling. But UNC immediately lost to Alabama in what was likely a critical result for many of our customer’s brackets.
While we also didn’t have any Alabama title brackets, the Tide bowed out to Connecticut at the right time, after knocking out a bigger chunk of the public.
We’ll also note that our four biggest public fades (i.e. popular teams not picked as champion on any of our customized brackets for standard scoring) were:
- No. 1 North Carolina (lost in Sweet 16)
- No. 2 Iowa State (lost in Sweet 16)
- No. 3 Kentucky (lost in First Round, coach promptly skipped town)
- No. 3 Illinois (lost in Elite Eight)
So by the time the Final Four arrived, there wasn’t a lot that could severely hurt us in terms of the public hitting a champion pick that our brackets totally missed.
4) The Longer Shot Champions Failed Spectacularly
While we noted several of the key aspects that likely led to higher win rates in smaller pool sizes, it certainly wasn’t all wine and roses for our brackets in 2024, because our recommended picks for larger pool size picks largely flamed out.
We had several teams pinpointed as underrated long shot champion picks for really large pools. They included BYU, Texas Tech, and New Mexico in some brackets, and those three teams all lost in the First Round.
No. 4 Auburn was also a key pick in some mid-to-large pool sizes, as an unpopular counter to the more common Connecticut champion run. That pick also went sideways in the First Round, when a series of unfortunate events led to Auburn getting upset by Yale.
The only long shot that did make noise in 2024, NC State, was largely absent from our brackets in terms of making a deep run. While that outcome should not really have hurt subscribers who played Connecticut-over-Purdue brackets, it did likely prevent subscribers from doing better in a lot of upset bonus pools, since our brackets were more likely to feature cinderella runs by other teams that did not happen.
Wrapping It Up
Playing in bracket contests is very risky business, especially when one or two game results not going your way can spell doom for your bracket (looking at you, Auburn). In any given year, even the smartest strategies can fail miserably.
Over the long term, picks that strike the best balance of risk and contrarian value based on your pool’s characteristics will generate the best returns. But even then, there are typically some very close calls to make for key picks like your NCAA champion, and the results of those close calls can spell the difference between winning and losing.
The 2024 NCAA champion ended up being the tournament favorite (well, the betting market favorite and our favorite, at least), and Connecticut became the first No. 1 overall seed to win in over a decade. The runner-up was also a “chalk” type of pick, as a fellow No. 1 seed.
Our best guess is that such an outcome will result in high win rates for our subscribers in smaller pools, but low win rates at larger pool sizes. In other words, we’ll likely see a good number of subscribers winning mostly modestly sized prizes (and bragging rights, of course, which are sometimes the biggest prize of all).
Last year was a bit of the reverse dynamic, with Connecticut’s win as an undervalued No. 4 seed leading to some nice paydays for our subscribers in larger pools, but lower win rates in smaller pools.
These past two years provide a good illustration of why we suggest playing a portfolio of brackets across different pool sizes and scoring systems each year. The smartest picks for differently sized pools are rarely the same, and if you cast a broader net, you’re more likely to catch a win in at least one pool.
Thanks And We Hope To See You Back
If you were an NCAA Bracket Picks subscriber in 2024, we appreciate your business. Our commitment to you is that we will keep working to improve our research and strategies every year, to deliver winning long term results.
In these types of pools, we expect the performance of our recommended picks to vary widely from year to year, and from subscriber to subscriber within a single year. We know it can be frustrating if you go multiple years without winning a pool, but that is the nature of this beast.
Since we started collecting the data in 2015, across thousands of real-world bracket pools, our subscribers have won prizes more than 3x as often as expected. We’re proud of that performance, and it shows that trusting the process eventually pays off.
Have a great summer and we hope to see you back for March Madness 2025!
Reminder: If you used the NCAA Bracket Picks product this year, please complete our subscriber survey if one is available to you, so we can collect real-world data on how our brackets performed.
P.S. If you want to read some of our past yearly bracket picks recaps, you can check them out at the links below: