Has anyone here heard of NumberFire. I use it to help analyze some of my matchups in fantasy football. I use this and what I read up each week to help determine which players to start. You can import your team and look at certain matchups. It uses mathematical forumlas and such to predict stats.

Here is a article about it http://techcrunch.com/2011/09/20/numberfire-raising-seed-round-to-build-the-last-fantasy-football-tool-youll-ever-need/

and here is the site http://www.numberfire.com/


This along with some other research can help fantasy teams out

Teams entered. I'll take all the help i can get considering i'm 1-5 overall across 3 leagues..

how accurate has it been for you? I've found most "prediction" sites are pretty faulty and only help you some with setting your lineups.

the prediction of yards and tds arent always right obviously but it helps when im comparing 2 out of 3 rbs or WR. Ill usually take the top two and since i dont have a very good def I sometimes pick a new one each week and it helps... I dont base all my decisions from numberfire but it is kinda cool look out and read some of their stuff.

I also think that since it uses mathematical analysis based off teams that it will probably be better as the season goes on. Once there are more games it might start being a better prediction it will have more games to go off of

Numberfire is suggesting I'll get more points from Darren Sproles this week than from Jahvid Best. Not sure about that one...

How the hell do you edit for your league scoring?

So far not real impressive.

The site lost all credibility when I noticed it’s using similar games dating back to 2000 to model production for this week.

I'm going to watch it for a while and see how it does in comparison to the ESPN rankers, and of course, texags posters.

quote:

---

ESPN rankers

---

Garbage!

Predicting with certainty TDs is stupid. You need to take into account change of a TD and weight that.

Hey guys,

Thanks for discussing numberFire. Wanted to make sure that we can address any questions you have. Feel free to reach out to us on Twitter or email at Connect@numberFire.com.

Some of the concerns I've seen relate to not being able to adjust for league scoring - the only response I have for that one is that we've got a pretty lean team right now and that it's actually a lot harder to take all the variables into account than you might think. It's definitely a thought on our minds, and we're going to be implementing that in the future.

As far as the actual comparisons go, we're running similarity scores that compare players to others who have played in the past in order to create similarity scores that give us a better understanding and ability to model future performance. The modeling that we conduct is very much cutting edge, statistically sound stuff. Our Chief Analyst has worked for ESPN and was on the cutting edge of what they were trying to do with data. You can find his posts and ideas about Markov Chains as they apply to football at drivebyfootball.blogspot.com.

Because of the way we do things, we're actually able to better predict player performance than the 'industry leaders' 70% of the time - meaning if we say Schaub will put up 20 points, and one of the leaders says 10 points, 7 times out of 10, we'll be closer to what actually happens. Also, for betting purposes, we are proven to consistently win people money.

Hope you guys find use in the service! Let us know if you have any feedback.

Almost all the prediction sites are garbage. I personally think walterfootball is the best one. Footballguys sucks. Never seen this one before might be alright. Yahoo and espn are constantly bad. I pay no attention to those sites unless I want to laugh at them.

quote:

---

it's actually a lot harder to take all the variables into account than you might think

---

Not sure I understand. To get to projected points you need to project TDs, yards, etc, right? How hard is it to change a TD from 6 points to 4 or 8?