Wednesday, 10 October 2012

I'm done

This project was fun while it lasted, but it just takes too much time to pull down and enter all the data every week. I feel like I can better spend that time on other things. I'm calling it quits for Power Rankings the Pros (unless someone is willing to pay me to keep going? no? no? Okay, didn't think so.)

Final thought: Home team wins more often than not.

I still wish that someone would develop a strong regression formula that would help predict the outcomes future match ups using offense and defensive statistics.

Enjoy the season.


Friday, 5 October 2012

Week 5 Power Rankings

Week 5 Power Rankings from 38 Sources.

Week 5 Picks

38 Power Rankings means 38 different predictions for this week's games.

I am a bit slow this week on the alternative rankings and alternative standings. I'll pick that up soon enough.




Methodology: 

Same as last week.

Tuesday, 2 October 2012

Week 4 Power Ranker Results


Methodology: Same as last week. In addition, I added a column to see how people did the previous  week so that I can track changes over time. 

The tiebreaker for the ranks after the percentage won is whoever was ahead in the prior week's rankings. The second tie-breaker is however the computer program decides.

Discussion: Everyone did really well last week. So, well in fact that 37 out of 38 teams are now doing better than coin flip for the entirety of the season. Perhaps the Power Rankings hold more weight as the season carries on? Or perhaps it was just a fluke week? We'll see.

No Where Plans is still #1 for the third week running. They did really well Power Ranking during the preseason and have been cruising on a 1 game lead ever since. CBS and Fox get another week at #2 and #3 respectively. 

That 38th team is Jeremiah over at NFL.com.

Plenty of Rankers went 13-3 over the course of last week. Well enough that I am not going to declare a winner for the week.

Friday, 28 September 2012

New Alternative Ranking - Home Team

If the home team doesn't win, its a shame.

Under someone's suggestion, I looked at what happened if you just predict that the home team wins every game. Record so far this season: 30-18. Winning percentage: .625; That is 1 game better than alphabetical order and 2 games better than the best power ranker. The sample size is still small, but there seems to be something to it.

Word of warning: Out of 16 games a week, the home team won 9 the first week, 14 the second week, and only 7 the third week. Some of it could have just been good luck Week 2, but it could equally be bad luck Week 3.

Also of note, if we take Week 1 out of the equation - because its really hard to rank teams during the pre-season. Then the Best Power Ranker is tied between Fox and CBS.

We will see how things change as the season progresses.

Thursday, 27 September 2012

Week 4 Picks

38 Power Rankings means 38 different predictions for this week's games.




Methodology: 

If the Power Ranker ranked the team higher this week, then they predict them to beat  out their lower ranked opponent.

Consensus is the team in which more of the Power Rankers had that team ranked higher than their opponent, thus predicting them to win.

Week 4 Power Rankings

Same 38 Leagues as last week. Same Methodology as last week. Same special thanks to reddit user skepticismissurvival.

 He added a bunch of these, then I added some more for week in and week out consistency. Click through the pages at the bottom to see comparison week over week. I am going to run some analysis now to see what we can find out...