Here’s Why 732,000 People Visited Our Site Last March Madness
Smart bracket pickers use TeamRankings.com because we deliver the most sophisticated, comprehensive, and objective bracket pick analysis.
We’ve been publishing data-driven bracket picks and advice since 2000 -- 18 years and counting. In the process, we turned a math geek’s hobby into a full-time business, with a team of engineers and quants dedicated to helping you win bracket pools more often.
This is not the kind of bracket picking advice you hear on sports radio. In fact, we often employ contrarian strategies that exploit irrational hype created by the mainstream sports media.
Bracket Picks Based On Science -- Not Personal Opinion
When we started TeamRankings.com from a Stanford dorm room, we were personally generating outstanding returns playing in bracket pools.
In short, we understood the science behind maximizing our odds to win a pool. We knew we needed to have accurate, unbiased predictions for any possible NCAA tournament matchup.
We also knew we needed to apply game theory principles, evaluating the risk vs. reward tradeoff of every possible pick given how our opponents were likely to pick their brackets.
However, collecting and processing the data we needed to do all this analysis took a Herculean effort.
The Data That Powers Our Bracket Picks
Every March, we’d spend hours scouring the web to gather the latest stats on teams.
We tracked the historical accuracy of different college basketball ratings and predictive systems from around the web, and incorporated the best ones into our research.
We monitored the betting markets. We pored over the latest player injury data. We used whatever sources we could to get a sense of teams likely to be overrated or underrated by our competitors in bracket pools.
With all that information in hand, an Excel spreadsheet, a lot of math, and an all-nighter or two yielded the main "calculated gambles" we’d make in our bracket contests each year.
Evolving Our Solution To Win More Pools
That process worked very well in terms of generating results. We were winning bracket pools much more often than a typical player would expect to win, but the research was exhausting.
As bona fide nerds, we had the math chops to manipulate large data sets and do complex analyses, but Excel has its limitations and the time pressure was intense.
As a result, we were still making some judgment calls or doing manual, "back of the envelope" type of calculations to make some pick decisions, especially when the numbers for Team A vs. Team B looked really close.
Although our approach to picking brackets was highly intelligent, there was plenty of opportunity to improve it. We wanted our process of optimizing bracket picks to be less hectic, and we wanted to be more confident that we had come up with the "right answer" -- the bracket that gave us the best possible chance to win.
So we challenged ourselves to build something better.
$150K Later: Technology-Powered Bracket Picking
Fast forward to 2018. Thanks to our growing business, over the last decade we have invested over $150,000 into bracket strategy research projects, technology development, and product improvements. It almost sounds silly, except to us it’s not. This is what we do for a living.
Today, we use the power of technology to analyze bracket pool picks with a level of precision that no human analyst can ever hope to match. As soon as the NCAA Selection Show finishes on Selection Sunday, our engineers go to work.
First, we aggregate thousands of data points on every 2018 tournament team, from power ratings to Vegas odds to public picking trends. Then, with the help of dozens of high-powered computer servers, we run millions of computer simulations of bracket contests.
(Fun fact: We spend about $10,000 every March renting server capacity from Amazon to do all our bracket related number crunching -- a bit more expensive than Microsoft Excel!)
Ready-To-Play Brackets, Customized For Your Pools
These computer simulations, and the related algorithms we’ve built, don’t just identify a list of the most underrated champion picks or promising upset picks for you to consider.
Rather, they exhaustively optimize every pick in the bracket, evaluating interdependencies and ensuring that the overall risk level of the picks is best aligned to your pool’s size and scoring system -- including dynamics like upset bonuses.
For example, if you’re competing against 10,000 people in a pool with big upset bonuses, the optimal bracket recommended by our system is going to look different than the optimal bracket for a 15-person pool with no upset bonus. In the first pool, picking a "15 seed over a 2 seed" upset in the first round may be a smart gamble; in the second pool, almost certainly not.
The good news is that although the scope of computation going on behind the scenes to optimize picks for your pool is massive, you get a very simple output: ready-to-play brackets, delivered in seconds.
The New Cutting Edge Of Bracket Picks
As time passes, more and more people are becoming familiar with the general principles of smart, value-driven bracket picking. With the data available today, it’s not super difficult to identify the tournament teams being overrated or underrated by the general public.
What remains incredibly difficult, though, is translating that general knowledge into 67 specific picks that give you the best possible chance to win your specific pool -- and having hard data that demonstrates why those picks give you the maximum edge.
To reach that level of confidence, you have to run millions of tests of potential bracket pick combinations, simulated in pools that look like yours. Anything less is an educated guess.
We’ve used data, technology, and math to build the new cutting edge in bracket picking. Today, the sophistication of our NCAA Bracket Picks product generates customized brackets that are up to 10 times as likely to win pools compared to the average competitor. And it’s simple to use.