Hello all!

I have added preliminary results from Lachine (top 25) to the true rankings site:
http://www.dragonboat-truerank.blogspot.com/

Of note, the weather clearly messed up times in a big way (times seemed to slow dramatically as the day wore on). For example, in the A knockout, times improved by 2 seconds after the completion of the first race. Thus PDBC A benefits big time. I have come up with a complex algorithem that could possibly remove this effect but I will let it be for now. Keep in mind that the data will improve as more races come in.

In order to help compensate for weather effects I did make one change. Because times earlier in the day were significantly faster than those at the time of the knockout, I averaged all knockout race times and counted these as two races. As a result, teams that lasted longer in the knock out did not get penalized for having a greater number of slower times. In all honestly, this changed things very very little.

Now that there is some data from both the West and East divisions, don't look for the true ranking lists to appear until TIDBF, Welland, (possibly Sudbury) for the West and Ottawa, Montreal, for the East. After this I will go about adding smaller festivals if I have time and there is enough statistical credibility to do so. Given the number of teams travelling around I should be able to add many.

Power Demon

You rock!

Power Demon, you've bent over backwards to include Lachine, but what about Stouffville? As a 500m race (approx), those results should have been much easier to include in your ranking.

Well, he did Lachine because he paddled there.

I don't think I agree with your logic about factoring in weather though. How would you factor in weather at an event that you didn't go to? You rankings are great....don't factor in the weather - it's too subjective.

maybe you should all read through his methodology before you complaining about what events ARE NOT included.

There are not enough teams at Stouffville who did pickering to do a good statiscal comparison. lachine was added to start an "Eastern" ranking. After Ottawa, TIDBRF and Sudbury their should be enough comparible data to start including festivals like Stouffville and Milton.

Power D, keep up the goood work.

Okay, lets look into that.

Teams in common Lachine vs Pickering:
- Verdun DBC - Adrenaline
- Pheonix ETS

Teams in common Pickering vs Stouffville
- Hanalei
- Scotia Rouge

Looks pretty even. Draw.


Distance at Lachine - 250m
Distance at Pickering - 500m
Distance at Stouffville - 500m

Hmm, Lachine is a short course, whereas Stouffville was a full 500m. Adv Stouffville.

You can't even compare Lachine to Pickering without doing some heavy duty extrapolation. I have time, but not that much time.

Scotia Rouge Times
Pickering - 2:07.53
Stouffville - 2:07.9

Hanalei Times
Pickering - 2:12.15
Stouffville - 2.07.3

Scotia is nearly identicle b/w Pickering and Stouffville. Hanalei seems to have gained some speed. But this should reflect in Hanalei's standings, and Scotia's time could serve as a benchmark for other non-Pickering teams that competed in Stouffville.

I think your missing something.

Lachine wasn't being compared to Pickering (correct me if I'm wrong Power D). Lachine had a lot of top Eastern teams there so it made sense to start an eastern ranking. A lot of those teams will be going to Ottawa and Montreal, and will "cross pollinate" with some of the western teams that are going to these festivals too.

There is not enough data to compare Lachine to Pickering as there isn't enough to compare Stouffville to Pickering YET. If Power D has enough time, I think you will find most festivals will eventually be included. Patience my little grasshopper.

The last poster was fully correct. There is enough data for Stouffville and Milton, in fact I have already built up the data for Stouffville. However, the true rankings are built up around two clusters the West (GTA) and East (Montreal). Lachine results are posted to give some perspective on how the East division crews look. I did the exact same thing last year.

Once more races are in, look for both West and East results to apprear. These two divisions are then joined and any smaller festivals added. It is not possible to join the two divisions until I have about 10-15 teams that have paddled in events across divisions(Pickering and Ottawa should help with this). We won't get a decent full ranking until a few more weeks.

As for the lachine adjustment, it would not have been added if I didn't think it was a good one and a general improvement on methodology. Here is an example that should illustrate what I have done. I will likely use this methodology for all knockouts.

Team A: race 1- 52.0 race 2-53.0 race 3-54.0 knockout 1-55.0 (knockout out in first race)

Team B: race 1- 51.5 race 2-52.5 race 3-53.5 knockout 1- 54.5 knockout 2- 54.5 knockout 3-54.5 (win's knockout)

Let's calculate two different averages:

METHOD 1: Straight Average
Team A (52.0 +53.0 +54.0 +55.0)/4 = 53.5
Team B (51.5 +52.5 +53.5 +54.5 +55.0 +55.5)/6 = 53.75

But how could this be when team B consistently beat team A? They are getting penealized for running more races during worse conditions. Because the number of races run by teams is uneven for a knockout we can improve the methodology.

METHOD 2: Equal Knockout Races
Here we make all teams run two knockout races.
Team A (52.0 +53.0 +54.0 +55.0 +55.0)/5 = 53.8
-average knockout time 55.0, put down two 55.0's
Team B (51.5 +52.5 +53.5 +55.0 +55.0)/5 = 53.5
-average knockout time 55.0, put down two 55.0's

Team B is now slightly faster than Team A as one would expect.

LOL! Ok, I just looked at the difference between the two approaches for Lachine. No team changes order and no team moves more than 0.2 seconds. This methodology improvement was probably not worth my time.

...oh well, here is to progress.

Cheers

Thanks, PD. Makes sense now. I should learn to read.

Interesting times re Hanalai. Sure looks on the surface like two very different teams. Maybe the Family Paul should get it's own ranking or something since Paul seems much faster than Hanalai without Paul.

What doesn't appear to be factored is that Pickering was gusting winds by the two finals, a better measurement would probably be the earlier times in the day.

Times in a water sport are not comparable from one race to the next, let alone one event to the next. Why the obsession? Whether it's in 1:54 or 2:44, or whether the winning time the Consolation Final would have been 6th in the Championship final, it's who finishes first that counts.

I guess you're not a math major, eh? PD uses a statistical analysis technique called regression analysis to allow us to compare results from different races. If all you care about is who came first, then you already have that - the race results - and you don't need to worry about that. If you want to know how two teams could compare to each other if they raced head to head (even if they never raced against each other all season), PD's rankings can give you some indication of what the outcome might be.