It would be nice to see some non ridicullous revised rankings that people might actually look at. Here's my suggested ranking method.

Firstly, to prove my point,
Huron Kayak with an average race result of 2 falls behind Miracle dragons with an average race result of of 38. This implies that Huron Kayak, if they had raced another race would come in 110th. This seems slightly less than likely.

I agree that boats that have proven themselves more should be ranked higher. Why not caclulatethe standard deviation from teams average finish, to their last place finish. Then apply this standard deviation to an individual teams average to calculate their third race.

This would provide a ranking that says, if a team did the additional race, and had their worse race of the year (as statistically expected), what would their ranking be.

This wouldn't be hard to do, like two columns in an excel book. Common.

WTF?

Ugh,...

I always get the people who flunked out of grade 10 math.

"Why not calculate the standard deviation from teams average finish, to their last place finish."

By this do you mean calculate one standard deviation based on all teams? or calculate N standard deviations one for each of the N teams?

"Then apply this standard deviation to an individual teams average to calculate their third race."

By 'apply' do you mean subtract the standard deviation from the average?

Maybe teams should race at least 3 times a year to be considered for ranking in the first place.

I am not aware of any teams that take Dragon Boating seriously and train at it regularly who do not do at least 3 races. Most do 4 or more.

Let teams like MCC and Huron who come out once a year to try and win flight tickets go on the Power rankings, but if they aren't racing enough to make the basic math work on the straight rankings then so be it.

Everyone knows where teams like MCC site in the overall scheme of things, but the fact is your aren't very likely to see them in the next regatta your entering.

Thanks for the intelligent question Floyd. I would advise using the standard deviation based on all teams such that you have the expected “worse race” possible in general.

When I say apply the standard deviation, I mean add it to the average race ranking under two races. This would be the race result of the third race. Then calculate the average again, using the two actual races and this hypothetical third race.

As for Spanky, the problem with demanding teams do many races is that qualifying races tend to only surround the Toronto area.

Qualifying Regattas: Wildwood, Pickering, Welland, TIDBRF-Corporate, TIDBRF-Community, Ottawa, Waterloo, Montreal, Summer Sensation, Woodstock, Stratford and Great White North Dragon Boat Challenge (911).

Therefore, it’s harder for Montreal crews (mainly) to do enough races to be considered. As a result, their festivals tend to have fewer top ranked crews, as a result Montreal local regattas get under represented in the selection pool. A better example than Huron would have been By-town (ranked right behind them). By-town would only have two festivals under a 4 ˝ hour drive. To ask them to do 4 is a bit much. With my ranking system, they would be harshly punished for not doing at least three races (by assuming their expected worse time possible), but would not be ludicrously punished by assuming they come in dead last in whatever they would have entered (after being second place for all other qualifying regattas). It still wouldn’t be a fair system, but at least it would have some viability.

Why bother? The rankings are like horoscopes...you might read them because they're available but they're still just the mindlessly, entertaining creation of an individual and have no bearing on the real world.

That's a pretty good point. I still figure that any team that is training any significant amount will get out there to some competitive events on a year.

Still, good point on the qualifying thing. I guess a team could race 5 times in Montreal and not hit 3 qualifying races.

To say that the straight rankings have no bearing on reality is just dumb. Clearly how you place in major festivals has some relation to how good a team is. By your reasoning, if a team won all festivals in Ontario, then clearly this would not be indicative of them being a good crew and could be fluke.

The real problem with the rankings is that they are based on rank when they should be based on time. An adjusted course speed should be created for each festival and average crew times compared across festivals. Ranking mean too little in terms of who are the fastest teams and no who chooses the right festival. If I have time I'll try to come up with something better.

No way should a ranking be based on time. Using that formula would give an advantage to all teams that race in good courses like Welland and Sudbury. The average time for all teams would be lower than that of Stratford or Toronto Island. For example, in the 2003 NorAm, MCC was able to post a 500m time that was south of 1:50. That same crew clocked 2:05 at TIDBRF the same year. Same team, different speeds due to different courses.

All rankings should be relative.

On teams placing against other teams in the rankings. And/or their festival placements should be weighted based on the strength of the festival. Festivals would then in turn be weighted by the strength of the teams that attended.

So similar to that GP Rankings guy does, but done scientifically by calculating the actual weighting of any given festival.

I figured out the formula for this a few years ago, but just never got around to cleaning it up formalizing the logic for all scenarios. For example if MCC entered and won the weakest festival on the planet the formula needs to make sure they don't get penalized for it. But if they somehow managed to finish below someone behind them in the rankings then they should take a bit of a hit along with the team beating them getting a bump up.

Anyway, as I said, I never polished it up to the point where it would be usable. While I think there is a degree of error in GP's method since he arbitrarily weights the festivals, it is still probably the most realistic ranking going right now. If he were to add in something where he only took the best 3 finishes for a given team (I can't remember if he did this or not already) it would be even better.

Cheers

The GP Rankings uses a teams best 5 point scoring races in calculating their overall score. This could mean that a higher placing at a smaller regatta could be replaced by lower placing at a larger, more competitive regatta if it scores more points.

The assignation of competitive multipliers is indeed arbitrary and imperfect, but helps to keep things simple.

will the ranking this year be kept up to date?

Clearly I did not mean take teams average times and there is your ranking. Times should be adjusted based on the speed of the festival. This could be done by taking times from all teams that compete across festivals and averaging them. For example, on average teams that compete in welland have a 5% lower time. Therefore, MCC's time in welland gets raised back up to account for the faster race course (or different boats, paddles, etc.). Once all festivals are standardized than an average time for all races would be calculated for a team. The fastest team ranks 1st.

The only problem I'm a little worried about is if the standard deviation of some festivals is smaller. For example, times scored in the Montreal basin seem really tight, the worst team seems unnaturally close to the first place teams. An adjustment for the standard deviation of the festival may be needed as well. However, if your just looking at top teams the bias (if it exists) should be small.

I think this is a much better appraoch as rankings don't mean too much. For example, Masawaska used to be nearly 20 seconds ahead of teams in Ottawa. Under a ranking system they would be placed first. If they were only 1 second ahead of the competition they would also be ranked first. Despite being a hell of a lot weaker team.