Playoff Bracket Simulation

[ebemiss]

Here are a couple bracket simulations. I used my overall ratings to decide the play-in games. Hit play again to run the tournament again.

Bracket scores only

http://nationalsportsrankings.com/joomla/index.php?option=com_tournament&magic=1403

box score and player stats (takes a couple seconds longer to load)

http://nationalsportsrankings.com/joomla/index.php?option=com_tournament&magic=1404

[UNHFootballAlum]

Very cool simulation

[tandemlax]

This is really interesting, even without having a clue as to what the inputs might be. I ran the simulation 5 times and saw a W&M- Delaware semifinal each time. Delaware won 4 of the 5 times, but never did the winning team score more than 15 points. Just curious if this was an abnormal result from a small sample size of if there is a systematic "good defense beats good offense" programming element? I'd expect that at least 20 points would be required to win this rematch, were it to materialize in a few weeks.

[larssc]

I ran it 25 times. I had played with it before the final selections and had liked the outcome. Similar results. I would assume more trials would give more probability of higher repeatibility.
Of course there is no way of knowing if the inputs reflect what will happen. Final wins: W&M 9, App.St. 7, VU 4, Del 3, EW 2.
Semis appearance: W&M 15, App.St. 11, VU 9, Del 5, EW 3, NH 3, Woff 2, SFA 1, CCU 1.
Should run with other first round winners in the set of 16 to be more fair.
Interesting exercise.

[cderry]

Well that's nice to just assume the winners of the opening rounds. I clicked refresh a bunch of times and Coastal always is in that spot over WIU.

This is pretty pointless.

[ASU Tailgaiteer]

Doesn't look to be very scientific. Why not just flip a coin or have a hen select games? Coastal got to the championship in my third selection. That's just not going to happen

[ebemiss]

just because your team doesn't win everytime, doesn't mean it's wrong or not scientific. Each teams has a rating and if you play the tournament thousands of times the "better" teams will win the majority of the time. My ratings are as good as any in the country for predicting the winning team in a game.

Don't assume just because you have Coastal Carolina advancing a few times it's going to happen. give me a break.

For the W. Illinois fan, i can set up the tournament with your team vs App st as well. I just used my ratings to fill the tournament of 16 as I only have a tournament format for 4,8,16,32,64 NOT 20.

[NHwildEcat]

I saw one simulation that scared the hell out of me...it had UNH losing to UNI again...ugh. What a nightmare!

[Wildcat80]

Just ran it 3 times---UNH beat del & wm BOTH times....lost to SFA in Frisco. I'm okay with that--this year!! GO CATS!!

[Mattymc727]

I ran it around 25 times or so and I couldnt even get UNH to Frisco.....

[bodoyle]

Well that's nice to just assume the winners of the opening rounds. I clicked refresh a bunch of times and Coastal always is in that spot over WIU.

This is pretty pointless.

Hey hey hey. I happen to like it very much ;)

[ASU Tailgaiteer]

just because your team doesn't win everytime, doesn't mean it's wrong or not scientific. Each teams has a rating and if you play the tournament thousands of times the "better" teams will win the majority of the time. My ratings are as good as any in the country for predicting the winning team in a game.

Don't assume just because you have Coastal Carolina advancing a few times it's going to happen. give me a break.

For the W. Illinois fan, i can set up the tournament with your team vs App st as well. I just used my ratings to fill the tournament of 16 as I only have a tournament format for 4,8,16,32,64 NOT 20.

I appreciate the effort xsmileyclapx, but unless you are simulating weather, crowd affect, etc....you may as well flip a coin.
Post your systems' best bracket, we can compare it to the actual games and if your 75% correct, I'll eat my words.
Oh, and if you simulate the first round as a round of 8, first round format (a tourmament within a tournament), it should take care of the non 20 snafu.

[profisme]

I tried over and over and over and couldn't get MSU past the semifinals. The closest I got was MSU playing Villanova in the semis and no result in the "winner" box. Delaware ended up playing nobody in the championship.

Then I clicked on refresh once more and in the box after Appy and Coastal Carolina, the winner turned out to be USC with no opponent in the quarterfinals. Then the winner of that quarterfinal game turned out to be Air Force. Talk about crazy.

[TheRevSFA]

haha first and third time i ran it..SFA won the championship beating Delaware both times.

Love it

[cderry]

I could create a FBS playoff simulator.



Oh wait. :D

[ebemiss]

I appreciate the effort xsmileyclapx, but unless you are simulating weather, crowd affect, etc....you may as well flip a coin.
Post your systems' best bracket, we can compare it to the actual games and if your 75% correct, I'll eat my words.
Oh, and if you simulate the first round as a round of 8, first round format (a tourmament within a tournament), it should take care of the non 20 snafu.

I will post the predictions for all the games if you'd like. I will post the first round games this week, the round of 16 before next week's games, etc... I'll also post the bracket using my numbers as they currently are, not adjusting them on a week to week basis. That will give us a projection for the whole tournament before it starts, and a week by week predictions with the adjusted numbers.

I'll also create the 8 team tourney for the play in games. I'll try to post it this afternoon, but I'm short on time and may have to post that and the predictions later tonight. thanks

Ed

[ebemiss]

Also, I'm not making any promises of the percentages it will predict correctly. It is just something I've created for users to have fun albeit with a statistically based engine.

That being said I'll glady try to reach your 75% goal. Here's a guy who used it a couple of weeks back and the computer actually beat his predictions for the week. He went 16-6, my numbers went 18-4.

http://www.mynewsadvance.com/index.php/chris_lang/comments/thursday_fcs_picks2/

Here's the link he posted before all of the games started, with my predicitions and his.

http://www.mynewsadvance.com/index.php/chris_lang/comments/thursday_fcs_picks1/

I understand my site and rankings are not perfect by any means, but it's better than a coin flip as I track them every week.