How flawed is the data poll?

[Matthew328]

Check this out..the data poll has Everman ranked 43 and Wylie at 12...yet Everman is favored by .5 points!! weird

[Titans]

Well Matthew:

If you look specifically at the web site, you will see that ONE person is in charged of figuring the "data" points.

With that in mind, I would say it means absolutely nothing and is worthless as Bill Clinton.

I would love to do away with all Top 10's/rankings/data polls, etc... in the high school ranks. Although it is good for bragging rights, it means diddly squat.

At the end of the season, when all the state champions are crowned (all 200 of them), I'll give my #1 ONLY. That's the only one that matters anyway.

Keep the faith!

[Matthew328]

Yeah I know it is one guy doing the poll...he does a good job his site is VERY informative..I thought thought it odd that he'd have Everman rated 31 spots lower than a team yet has em favored...

[sinton66]

He may even be using a different formula than last year. I thought I saw something about that earlier. What he has on there now seems to be mostly relative power rankings. The MR thing is something he got off the net about handicapping games, and I think he is just trying it out against his own method. Last year he leaned very heavy on pf/pa ratio as I do. He has changed something, I just don't have a handle on it yet. Starting with regional finals and going on up, the pf/pa ratio comes heavily into play. From that point on, strength of schedule arguments go away. The numbers themselves will show adjustments in that department. If your team is presently at least at 3 to 1, you have a shot. Lower than that, I wouldn't expect too much.

[29x281]

The poll is as valuable as any other (not valuable at all). But the way the site is broken down makes it far and away the most convenient site around for researching Texas high school football. I think we can all live with the poll as long as the format of the site continues to provide information as efficiently as it does. I'm glad we have it.

It seems to me the poll is a rating of achievement, not ability. Having Burnet #1 is more a reflection of Burnet having achieved more during this season SO FAR than any other team. Everman may be better than anyone else around, but they haven't achieved much so far this year.

Adding to that is the fact that the win over Lawton Ike isn't figured in, so to the computer Everman lost against the only good team they played. It's understandable that they're ranked in the 40s.

[3ABirdMan]

I am not going to pretend to be a mathmetician, but I believe you have to look at how the ratings are calculated and try to understand how the formulas work.

Any game "pick" is figured on the difference in the Power Rating number of the 2 teams, not neccessarily what "place" they are given in the order from top to bottom. New Power Ratings for each week are calculated backwards from the last game played to the first, recalculated, and then from the first to the last, recalcuated, and back and forth until these new calculations don't change anything, meaning the numbers are close to "true" and free of error. One MAIN POINT to remember is all of these calculations are based on ALL TEAMS PERFORMANCE, not just one. This tells me that there is a mathematical function being generated and applied to ALL TEAMS, top to bottom. It also tells me that this function is not linear, meaning the difference in teams #1 and #10 is alot less than the difference in teams # 50 and #60 ( the closer you get to the top, the more comparable the teams are - ALL GOOD).

In the case of Everman, you must look at the "Blow-out Factor" that is applied. 7 of Everman's 10 games have been won by more than 40 points. This actually hurts Everman compared to other teams, because of the point spread. The margin of victory is rated by the number of points in the victory, with the 1st point being the most important, and the last point being the least important. In other words, winning against a viable opponent by more than the difference in the Power Ratings of the 2 teams actually begins to hurt your rating. Fair? I don't know, but it does allow for blowouts. It also explains why Everman is a lot lower than everyone thinks they should be. Why are they favored over Wylie when rated lower? I said at the top, I'm not a mathematian - I'm not sure, and not defending the pick either.

I posted on another thread that I realy do like the mathematical ratings vs the "emotion" of some polls, like the AP. I also believe you have to read and have a basic understanding of HOW the numbers are calculated in order to accept or reject them. The thing I don't like about Collier-Sharp and Texas Power Ratings is they don't explain their calculations, and DataPoll does. That's why I also liked Freeze Ratings last year, but he didn't do them this year.

These things should be an aid to us in our discussion, but you have to put some effort in trying to learn why teams are where they are, not just criticize because we don't like the numbers.

[PPHSfan]

Matthew,

3ABirdman did an excellent job of explaining why Everman is not ranked as High in the Poll.

Now I am going to tell you why they are still favored.

Everman has a power rating of 112.0 and Wylie has a power rating of 111.9

It's just that simple :D :D

[Matthew328]

Well a math major I am not...it just confused me when I saw Everman at 43 and Wylie and 12 and Everman was still favored..thanx for explaining it guys!

[sinton66]

I'm not a mathematical genius either, that's WHY I opt out for a simpler method. Mine makes it necessary to use my "gut" more, but it seems to work for me. I really like Data Poll because as said, you can't get the info anywhere else, and he does at least attempt to explain his method(even though it tends to sail right over my head). He does a very good job of keeping the Data up to date on a timely manner, a fact I truly appreciate.

[Rabbit'93]

Matthew haven't you ever ranked a team and then picked the lower seed to win. All polls (especially HS) are inaccurate. There is no way to compare apples to apples, so to speak.

[Matthew328]

Yes I have as a matter of fact...LOL but not 31 spots different!! :cool:

[whtfbplaya]

All polls are flawed.

[Old Cardinal]

Matt, I think my occasional "Poll not from the 6 DFW Pollsters" has proven uncannyily accurate. The leading criteria being, if all things are comparable-go with the larger enrollments and dropdowns. It works in both Div. I and Div. II. On that basis alone, big ole Jasper ought to whip little ole Cameron Yoe...Now that scares me, Elgin is the largest 3a enrollment in the state, at present,200 more students than Bridge City. The balancing factor(I hope)is that BC has some 3rd year and 2nd year playoff veterans that have a lot of "field savvy".