Men's National Championship
#2 Villanova 77
#1 North Carolina 74
2.9% picked the winner of this game
54 of 63 correct in best of 113316 bracketsWomen's National Championship
#1 Connecticut 82
#4 Syracuse 51
54% picked the winner of this game
49 of 63 correct in best of 420 brackets
The Pick65 Power Ratings assigns a rating for each Division I men's U.S. college basketball team, and ranks them based on that rating. A higher numerical rating indicates a stronger team.
The Pick65 Power Ratings are formed by analyzing scores for all games played during the season. Actual win/loss results are not considered, only points scored and allowed and the strength of the opponent. Furthermore, no points scored in overtime are even considered. The point difference against each opponent is measured against how other teams fared against the opponent to arrive at an overall power rating. The resulting power ratings are sorted from top to bottom to form the ordered list.
While these ratings have merit, ultimately the best it can hope to be is an approximation of teams relative strength. The power rating does not take into account the fact that non competitive games (i.e. blow-outs) often give false impressions of relative strengths since both the winner and loser loses incentive to play their best and use their best players. Also, this system does not take into account injuries or general upward or downward trends, but rather utilizes all games played independent of how recently they were played.
No. The players on the court are trying to win each game and are not expected to be concerned with the margin of victory or defeat. For this reason, it would not be fair to use a "power rating" to judge worthiness of teams. However, this system can be useful as a PREDICTOR of the outcome of upcoming games. The system is not meant to be used to dole out rewards such as post season tournament berths to higher rated teams.
Yes. The home court advantage is measured based on a subset of games played so far. In general, the home court is worth about 6% more points, or 4 points on an absolute scale, but that can vary as new results are considered. In measuring the home court advantage, we are selective in which games we count due to scheduling biases.
This question usually comes to mind when "team A" is much lower in the human polls. Computer rankings have advantages and disadvantages over human polls. First of all, it is difficult for humans to consider margin of victory and strength of opponents for many games, while a computer can exhaustively examine every game. Secondly, traditional powers such as Kansas and Duke might have an advantage until actual loses prove otherwise since they are usually strong year after year. A computer can notice when a team is winning by lower margins of victory than expected and can can distinguish a win over a strong team from a win over an over-rated team.
Human polls on the other hand can take in much more intangibles factors. Some computer ratings may take into account upward and downward trends in ranking the teams, however as of now the PickHoops Power Ratings do not take this into account.
Ultimately, if a team is rated higher than expected, it can mean they are somewhat of a dark-horse, or the computer has over-ranked them, or even a combination of the two. A similar statement can be made about teams that are ranked lower than expected.
No one would deny that all teams have good days and bad days. So how do we know that when "team B" beat "team A" that "team A" was not just having a bad day? For example, in a recent year Purdue was the lone team to have beaten Duke late in the season, but few would suggest that Purdue was stronger than Duke. This question rarely comes up when there is a wide gap between teams. But if we accept the fact that Duke should be rated higher than Purdue, even though Purdue won head to head, does it not make sense that "team A" could be rated only slightly above "team B" even though "team B" beat "team A"?
The power rating is the ratio of expected points scored versus expected points allowed against a mythical "average" team. Thus roughly half the teams will have a power rating above 1 and half below.
We try to exercise reasonable care in recording game results, however since all PickHoops contributors have demanding day jobs and families, some errors may have slipped through the cracks. We welcome any correctional feedback that does not call into question the matrimonial status of our parents at the time of our births. E-mail that has correctional information should ideally have a web link to a teams complete record.
The PickHoops Power Ratings has had reasonable success in picking game outcomes. To get a rough predictor of the final score at the end of regulation of an upcoming game, take the average point total of 65.4 and multiply by the power rating of each time. Multiply the home team, if any, by 1.029. This offers only a crude predictor, which doesn't take into account offensive-defensive tendencies. We do NOT encourage using these calculations as a gambling resource.