(03-07-2019 09:11 PM)UofToledoFans Wrote: (03-07-2019 08:52 PM)emu steve Wrote: (03-07-2019 11:36 AM)UofToledoFans Wrote: (03-07-2019 01:33 AM)MaddDawgz02 Wrote: (03-06-2019 11:10 AM)T4C8 Wrote: As a Toledo fan, and who I think we are best/worst against, this is an absolutely perfect bracket.
I agree. Right now I have Toledo at its highest percentage to win the MAC tourney that I have had all year long (22%), and I think thats because the projected bracket lines up favorably for them
I'm not good at math but if Toledo plays three games against even opponents our percentage to win all three would be half of half of half... So 12.5%. what's your percentage for Toledo to beat UB? I'd have it at like 22%. That can't be the chance to win the league though unless we had 100% chance to win the first two. There is no one I am 80% certain we beat but WMU and Ohio. Most of the league feels like a 60-40% because of how deep it is.
Yes, if a team has a 50% chance of winning each of 3 games, the probability is .5 x .5 x .5. Or .125.
It is the same as flipping a coin 3 times and getting 3 heads.
He obviously has us heavily favored in the first two games which I disagree with. That or he is giving Toledo lots of credit in a possible game vs. UB.
Correct. Here is how I break it down. Toledo a 2 seed, I look at the 10 vs 7. For the 10 and 7 team I calculate the probability of each winning that game. Let's say for example, it is 10 NIU vs 7 Akron. For Toledo's first game I use the rankings I have for each NIU and Akron and build an opponent for Toledo based on those probabilities. Taking NIU vs Akron right now for a game in Akron I get Akron with a 68% win probability. Then Toledo's opponent and probability to win their first game would be against a team built from taking the power rankings from a single opponent made up of 68% Akron and 32% NIU. Essentially I am creating a fake opponent, the Akron Huskies if you will. So it is hedging Toledo will play Akron in Round 2, but offering Toledo slightly higher odds than playing Akron straight up since the weaker NIU has a 32% chance to win. This process is repeated all the way to the finals. Here are the individual round by round percentages I have for Toledo right now:
Quarterfinal game: 83% to win
Semifinal game: 78% to win
Finals: 34% chance to win
The finals opponent using the above methodology would be heavily weighted to Buffalo obviously, but includes slim chances it is not Buffalo (currently have it 85% weighted to it being Buffalo)
Multiplying those individual game percentages above gives the 22%
The two major shortcomings of this are obvious:
1) Coming up with the initial power rankings to get the probabilities of individual games, obviously nothing is perfect and my rankings are flawed and
2) This idea really only has meaning if the brackets are set already. The calculation obviously comes way too hard for me to think about when you have to think about what are the chances NIU finishes as a 10 seed and Akron as a 7. For simplicity, one has to assume the seeding forecast is perfect (which it will be in 24 hours hah).