Looking at the Data: Coyne builds solid lead heading into Sweet 16

by: The Cold Hard Truth

After the first two rounds, we've all picked 48 games. If you are in a March Madness pool this year and have successfully selected the winner in three-quarters of those games, you're probably close to the top, if not leading your competition. Heading into the Round of 16, Mike Coyne from Mansfield, Texas has a commanding 83 point CBCC. At this point, he has correctly chosen 36 games. So what's his secret?

Okay, I don't know Mike's secret method of picking games, but I can tell you that I'm a "gut feeling" kind of guy. Using that method, I've slip into the Top 10 on several occasions in CBCC history, once ending up in the #2 position. But let's face it, no one remembers second place. So this year, I tried something different - I looked at the data. I evaluated the following: Assist to turnover ratio, rebounding, defensive efficiency, and offensive efficiency. My hope was to find the secret method of selection that would take me to the top of the heap.

Looking across the board at all four categories, guess which team had the best data? Michigan State. So I picked them to win it all, just like many of you did. Clearly, there are some things that data simply can't predict. I wasn't going to cling to Middle Tennessee's #172 offensive efficiency and come to any conclusion that gave them a chance to win. It just happens sometimes - that's why we call it "Madness". So I graciously accept the "1" confidence rating I placed on Middle Tennessee and have moved on with my life and the rest of my bracket.

So if you picked all the higher ranked teams, you got 32 games right, including a mediocre 19-13 first round. If you opted to take all the teams with better ball handling skills (i.e. better assist to turnover ratios), then you went a decent 24-8 the first round, and a heart-wrenching 6-10 in the second round for a combined 30 wins. Another option - we always hear people say that better rebounding teams win more games. Picking those teams would have produced a losing record and only 21 wins through two rounds. What about defense? It's the best defensive teams that win games - right? If you picked those teams, you had a record of 25-23 for the first two rounds. My conclusion, you can't just consider rebounding and defense without taking a deep look at the competition, and to do that, takes way too much time.

So that leaves offensive efficency as our final predictive factor I looked at. If you selected all the teams that run really good offenses; teams that can shoot free throws really well, and can do all the little things right that lead to scoring - if you picked these teams, you went an incredible 39-9! That is 25-7 the first round and 14-2 in Round 2. You would have lost the Notre Dame/SFA on a tip in with one second left. As a side note, each of these teams played great on offense that game - a combined 56% shooting in the first half. Their offensive efficency rankings for the year were the best combined totals of any teams playing second round games. They certainly delivered as advertised, and SFA could have easily won that game. The other second round game you would have lost - Wisconsin hits the incredible buzzer beater over a better offensive team in Xavier. This was another game that could have easily gone the other direction. So in this case, a perfect Round 2 was almost within reach. Clearly, if we had used this method of game selection, Mike Coyne would have some competition. As it is, we all have an uphill struggle to catch our current Round 2 CBCC leader.

Another interesting offensive stat I noticed as I carefully evaluated these games: Each team that scored more points than its opponent won their game. This resulted in an impressive 48-0 record through two rounds. This might suggest that we choose our winners after they play the games. It would be so much easier.

As another great weekend of basketball approaches, there is only one question that remains unanswered. How did Northern Iowa lose that game?