karl(east) wrote:scorekeeper wrote:karl(east) wrote:
What do you mean by "actual" rankings? AP/Let's Play Hockey/MNHockeyHub are all the same thing, and they're already putting out rankings.
I mean fact based rankings. Mathematical.
The ones you mention are all great fun, but they are just opinion polls. They are a great conversation piece, and a lot of fun to debate but in terms of accurate quantification of team strength, they have inherent flaws and are notoriously inferior to mathematical quanitifcation.
Of course, mathematically, it's impossible to have a pre-season poll, so there is definitely a place for opinion polls in that regard, but once the players take the ice and the games are being players, results replace opinions.
Alright, makes sense--no doubt the math-based ones have some major benefits that us silly humans don't have. Within the confines of what they measure, they are 100% correct. The problem comes from those confines, which are selected by a human somewhere along the line.
Strong mathematical rankings are the simplest thing in the world. It's very simply
margin of victory over strength of schedule. There is no human element to it whatsoever. This has been tried and true in Las Vegas for decades and billions and billions of dollars have been made pitting math against humans.
The problem comes when humans try and mess with the math. The BCS is a perfect example. The original formula was perfect, but then got messed with as humans wanted more influence over the results. Adding in coaches and AP for example made it worse, not better.
Las Vegas uses a strict
margin of victory over strength of schedule. Adjustments are only made for weather and injuries to key players, and even these are very rare.
It's not uncommon to find football teams higher ranked in the AP, for example, as underdogs to lower ranked teams and even unranked teams. While opinion polls are fun, you can actually book big money - and it's done everyday - on mathematical polls.
MYHockeyRankings.com has been notoriously accurate - MORE accurate on Minnesota hockey than experts from right within the state.
The MYHockey algarythm was written by a man in Indiana who likely has never seen a Minnesota team play yet his algarythm is decidedly more accurate than the local experts watching these games first hand.
I believe the sole flaw in the MyHockeyRankings is in the cap at 7 goals. I believe the cap should no less than the initial margin between the two teams.
For example, if a team with a Rating of 14 plays a team with a Rating of 2, the highest Rating the better team can get for the win is 9. Even on a score of 16-1, for example. I believe they should at least be able to recoup their own initial Rating (in this case of 14). Other than that, the MyHockeyRanking system is a mirror of the oddsmaking systems in Las Vegas.
I do keep a personal set that doesn't limit the margin, but as rare as blowouts are, the MyHockeyRankings set is the best public ranking system available for Minnesota Hockey teams.
Anyways, probably hi-jacked this thred enough. It'll be an interesting comparison on December 12th. I think the inital set comes out December 5th, but there won't be enough data to get an accurate read. You really need about 5 scores on each team and good crossover between conferences to get an accurate read.