<< Previous Topic | Next Topic >>Return to Index  

Rankings update

September 16 2009 at 10:20 PM
MJC 

 
Just a quick note that the GWN Challenge results have been added to the rankings:
http://www.mofosdragonboat.com/Rankings/RankingsIndex.php

All 500m races including industry cups were factored in to the rankings. Remember, the ranking uses adjusted race times but you can always see your team's unadjusted race times in the race archive section.

We will add one more regatta (the Quebec Cup) before calling it a wrap for 2009. Thanks to everyone who sent suggestions, and pointed out problems. Please keep them coming. Special thanks to Sara for getting the idea off the ground and helping with the website, Peter for helping with the data, and Power Demon for laying out the framework. We plan to keep the ball rolling in 2010.



-Matt

 
 Respond to this message   
AuthorReply
Anonymous

Re: Rankings update

September 16 2009, 10:31 PM 

Great job MJC with the rankings!

 
 
Anonymous

Re: Rankings update

September 16 2009, 10:52 PM 

Why are the industry cup races included...since only a small percentage of teams participated in those? I don't think that they should be included...just my opinion.

 
 
Anonymous

Re: Rankings update

September 16 2009, 11:03 PM 

So the Mofos would be ranked higher perhaps?

 
 
Anonymous

Re: Rankings update

September 16 2009, 11:29 PM 

great work MJC, keep it up looking forward to next year

 
 
Anonymous

Thanks, Matt.

September 17 2009, 8:23 AM 

Love the rankings, but I agree that challenge cups should not be included. The reason being that our team only races these for the win, and not for time, so we take the foot off the gas when the lead is enough to be certain. Other teams may do the same, which would make them appear slower than they really are.

Also, GWN this year had incredible wind variations. It went from light headwind, to strong headwind, to strong tailwind. The same teams were logging anywhere from 2:02-2:04's in the first heat on Saturday morning only to see times go up to 2:07-2:11 in the afternoon, and then back down to 2:00's again on Sunday (even sub 2:00 in Hydro's case).

I'm not saying you should throw the results out completely...but how can you include them when they vary so wildly?

 
 
Anonymous

Re: Rankings update

September 17 2009, 8:52 AM 


"I'm not saying you should throw the results out completely...but how can you include them when they vary so wildly?"

It all comes out in the wash - so what if GWN scews your ranking time by like 0.2 of a second.

Most teams have logged at least 10+ races this year and sometimes you get the faster lane some times the slower. At the end of the day the rankings will give a good rough approximation of where your team fits in - and if in doubt you can even use the Mofos race comparison tool to see if you tended to beat the crews ahead of you head to head.

 
 
Anonymous

What I'm saying...

September 17 2009, 9:02 AM 

Is that for teams that competed at GWN their times would have been overly inflated by Saturday afternoons big headwind, relative to teams that did not compete at GWN. You could make a similar, but less strong argument about Sunday afternoons times being too fast, but the effect was much less noticable.

I'm not worried about my own teams ranking. A quick check shows me it didn't move anyhow, but even so that's not the point.

Consensus on Saturday was that afternoon times were about +4 to +6 seconds slower. That's a HUGE variation. Could it be smoothed down?

 
 
Anonymous

Re: Rankings update

September 17 2009, 9:14 AM 

He normalizes the data. Read the methodology.

 
 
Anonymous

Re: Rankings update

September 17 2009, 10:35 AM 

Re: race to win vs. race for time

Shouldn't these be the same thing? I mean, if you have optimized your racing strategy for time, would you also not have optimized your chances of winning?

The only case where this is not true is if you're so far ahead of other teams in the race that you don't paddle as hard. But that rarely happens. And if a team does that, well, they deserve their rankings to be pulled down. I'm not accusing your team of doing that, of course.

 
 
Anonymous

Re: Rankings update

September 17 2009, 10:41 AM 

Except with the industry cups, some were hotley contested with teams of similar abilities, while others just needed their regular start and then coast to the line.

I see an argument for both inclusion and exclusion, but there definitely were a few teams in the industry cup challenges that just coasted as there was no challenge.

 
 
Anonymous

Re: Rankings update

September 17 2009, 10:50 AM 

Why is Stratford not being used in the rankings?

Correct me if I'm wrong, but with teams like Hanalei, The Blades, BMO, Golden Plate, etc. all going, there should be enough data to add it to the rankings as well. I mean, if Pickering HS, Welland, Waterloo, Hamilton, and Oakville are in there, what reason would there be to not include this one?

Thank you.

 
 
Anonymous

Re: Rankings update

September 17 2009, 11:01 AM 

Well Hanalei isn't being called "Hanalei" for this festival. So unless you have first hand knowledge of the team members it be hard for Matt to put them as Hanalei.

Lots of teams pack it in after GWN, and in many cases Stratford is a case of filling the boat, maybe a not true reflection of the team? Of course everyone knows about the wonderful lane issues at this course.

 
 
Anonymous

Re: Rankings update

September 17 2009, 11:51 AM 

Industry cup: to include or not?

If the rankings start excluding certain races, then it'll get ugly fast!

What if there was a HUGE tail wind in the afternoon and all of the industry cup goers have significantly faster times? I doubt there would be this much whining about this topic! Also, should the rankings exclude those teams that raced early enough to squeeze in 2 races before the headwind came in?

Bottom line is that there are conditions that can not be controlled but the rankings will normalize to show an accurate time. One race out of 10+ won't make a huge difference.

 
 
Anonymous

Re: Rankings update

September 17 2009, 12:16 PM 

RE: Previous Stratford post (2 posts ago):

Your post was not fair at all. If a team has "packed it in" then they wouldn't waste their money doing more regattas. I assure you that Golden Plate, The Blades, BMO, and Hanalei will all bring similar rosters as other regattas. Your season might end at GWN but theirs doesn't and therefore they are still practicing and preparing to do as well as they possibly can in Stratford.

As for Hanalei, yes it is them. Ask any of them and they'll tell you this. They change their name for this regatta in honor of one of their paddler's father - a very classy move in their part. In fact, I believe they donated to charity their $500 winnings from last year's 100m sprint victory also in Lloyd's memory. This is not the first year they do this (the name change).

By your rationale then any team that has a bad showing at a regatta could ask to have that regatta removed from the rankings because "it wasn't their regular roster". I believe that as long as you enetered your team in a regatta it should count towards the rankings (with the possible exception of Hanalei who registered under a different name and could therefore challenge this regatta being included with their other ones - something I doubt they'd do). All the other teams, though... tough luck!

 
 
Anonymous

Re: Rankings update

September 17 2009, 12:18 PM 

... and there are lane issues at every course, so that shouldn't affect the ability to include Stratford in the rankings either. Are you telling me that Montreal, Toronto Island, GWN, etc all have even lanes?

 
 
Anonymous

Re: Rankings update

September 17 2009, 12:32 PM 

Geez dude, relax.

I was hypothizing why Stratford might not be included in the rankings. It seems you have a vested interest in the rankings or something. Its just a fun thing, I could care less if Stratford is included or not.

 
 
Anonymous

Re: Rankings update

September 17 2009, 12:42 PM 

"By your rationale then any team that has a bad showing at a regatta could ask to have that regatta removed from the rankings because "it wasn't their regular roster". I believe that as long as you enetered your team in a regatta it should count towards the rankings (with the possible exception of Hanalei who registered under a different name and could therefore challenge this regatta being included with their other ones - something I doubt they'd do). All the other teams, though... tough luck"

I think I would reverse that. If a team races under another name, regardless of the reason, they should be listed under that new name. However, if they request that those races be added to their regular name, that's when you do so.

 
 
Anonymous

Re: Rankings update

September 17 2009, 12:55 PM 

That's exactly what I meant... sorry if it wasn't clear. In other words, if they chose to, Hanalei would be the only team that could have these races not added with the others (because of the different name). The other teams, though, wouldn't have that option.

 
 
MJC

Re: Rankings update

September 17 2009, 7:15 PM 

OK, it looks like there are a decent number of teams in common between Stratford and the rankings so we should be able to include Stratford. We will take a careful look when the results are out. At the very least we will add Stratford to the race archive.

My personal preference is to include the industry cup races simply because they add more data points and it keeps things simple: include all races. For most teams the industry cup races are competitive. Power Demon, if you are out there did you include all races as well?

Thanks to the people who emailed us regarding slightly different team names that should be treated as the same team. We updated the GWN team names so that now more teams are recognized as being in common between GWN and the rankings. This produces a slightly different adjustment curve and most team's GWN times shifted by a couple tenths of a second as a result.

 
 
Anonymous

Re: Rankings update

September 17 2009, 11:30 PM 

MJC, I know it's been said before, but thank you for what you're doing with this year's rankings and your commitment to make them as fair as possible. I hope you continue this next year and in years to come.

As much as it is important to keep the relevance of the rankings in perspective, it is a very useful tool to track your team's progress through the year and against other teams.


 
 
Anonymous

Re: Rankings update

September 17 2009, 11:58 PM 

I move that you exclude the Saints' results from the last TWO regattas. Or simply name them Saints (Pre July) and Saints (Post July) to reflect the noticeably different rosters.

 
 
Anonymous

Re: Rankings update

September 18 2009, 12:03 AM 

I move to ignore idiots passing motions and only seperate the Saints into two teams IF the Saints request it themselves.

 
 
Anonymous

Re: Rankings update

September 18 2009, 12:27 AM 

Of course they wouldn't request it themselves. It's only beneficial for them to do what they did. That's why everyone's so up in arms about it.

 
 
Anonymous

Re: Rankings update

September 18 2009, 8:09 AM 

You might as well just stop whining about this type of thing. There are no "rules" as to how your festival crew is set up.

Unless the Saints manage to keep all these paddlers next year then they will drop off again.

Another example would be Hanalei at Nationals. This was the "Club Crew" version of their team and included a significant amount of CSDC Seniors. This is actually a team we most likely will see from time to time, but it surely isn't the Hanalie that races all the other times.

We will most likely see this more and more as these larger Clubs/Alliances form so we might as well just get used to it. Any attempt by the rankings folks to keep track of things to that level would be a complete nightmare.

I don't see why people are oh so hung up on how another team ranks. Heck, if GWN moved The Saints above your team then just pretend it didn't and mentally move yourselves up a spot. It's like the guy freaking that Stratford will not be included. Hello, having Stratford included will not be improving anyones standing no matter how much adjusting is done.

 
 
Anonymous

Re: Rankings update

September 18 2009, 8:31 AM 

Why are we giving 'The Saints' so much publicity?

Kanye version of dboat? anyone?

 
 
Anonymous

Re: Rankings update

September 18 2009, 9:02 AM 


 
 
Anonymous

Re: Rankings update

September 18 2009, 6:08 PM 

MJC, just wondering. How do you decide which team is seeded higher when more than one team end up averaging the same time? For example, I notice that there are three teams at 2:11:3 (RBC Comets, Victoria College, and PDBC U23 I). I would have thought that Victoria College would be ranked above the other two simply because they have more regattas included in their rankings (the other two teams only have one regatta included in their rankings while Victoria College has three), making their average time more "reliable".

Is there another variable you are using, like having more decimal places than what you are posting? Or maybe there's another "tiebreaker" you are using (it doesn't appear to be alphabetical).

I used this only as an example, the same scenario plays out for other set of teams (see 2:04:2, 2:07:4, 2:07:9, 2:08:1, 2:09:1, 2:12:3, etc.). In all these cases there are teams that have more sample regattas seeded under a team with fewer.

Thanks.

 
 
Anonymous

Re: Rankings update

September 18 2009, 6:18 PM 

I believe there's another significant digit involved.

 
 
Anonymous

Re: Rankings update

September 18 2009, 6:23 PM 

Even if you ignore possible rostering or gear downs in the Challenge Cups, you cannot change the fact that any team doing one got a 2nd race into the headwind to effect their overall times.

It's like averaging out another 5-7 seconds across all the races they did. I bet if you were to remove the challenge races for the teams that did it, their average time would drop by 1.5-2s each. That is more than is acceptable if you want this to be considered even for all teams.

 
 
Anonymous

Re: Rankings update

September 18 2009, 6:33 PM 

Not 100% true.

Look at Komodo Dragons, GSK Fire Dragons, Riptide, Iron Dragons, York Seawolves, RBC Royal Dragons .... All of these team's industry cup race times were faster than their 2nd race (some even faster than their 1st heat in the regular conditions!!!)

How do you handle that situation then?

 
 
Anonymous

Re: Rankings update

September 18 2009, 6:38 PM 

Easy, eliminate all the Challenge cups from the rankings. Obviously some wonky stuff going on there if that is the case.

 
 
Anonymous

Re: Rankings update

September 18 2009, 6:44 PM 

How about we only include races that you want to be included. Maybe Matt can set up a login for every team so that they can choose which race they feel should be included?

 
 
Anonymous

Re: Rankings update

September 18 2009, 7:04 PM 

yeah now we're talking...oh oh I only want the race that our team did under 2:00! what do you know we're just behind the Preds

 
 
anonymous

Re: Rankings update

September 18 2009, 11:02 PM 

Why is there not an update of time for the Entre Amis women? They raced at GWN?

 
 
Anonymous

Re: Rankings update

September 18 2009, 11:41 PM 

They're 16th in the Women's rankings.

 
 
Anonymous

Re: Rankings update

September 19 2009, 1:49 AM 

Was the Entre Amis crew from MIDBRF different than the one that raced at GWN? I assume that is what the previous poster was asking.

 
 
anonymous

Re: Rankings update

September 19 2009, 3:37 PM 

Entre Amis ladies crew is the same crew that raced at Montreal...the Hydro and Jetstart ladies.

 
 
Anonymous

Simply a joke

September 20 2009, 12:44 AM 

I completely don't understand how these times could include challenge cups to which most teams (as stated before) go merely for rank and often aren't trying their hardest to win. That purely negatively impacts teams that decided to do and take it easy on the extra races.

Further, you ranking scheme is skewed. Have you included any weather components? Have you normalized for differences in the courses? Did you even consider teams who prefer to ditch their team's paddlers for stronger paddlers to stack their boats? Seemingly this is becoming a more prevalent item. If you've got part of your boat filled with Preds you're obviously going to fair quite better than with you're original team.

Don't post rankings with if you are going to half-ass the measurement criteria. It just cheapens the sport.

 
 
Anonymous

Re: Rankings update

September 20 2009, 2:05 AM 

if you dont like it - do your own.

 
 
Anonymous

Re: Rankings update

September 20 2009, 3:11 PM 

"Further, you ranking scheme is skewed. Have you included any weather components? Have you normalized for differences in the courses? Did you even consider teams who prefer to ditch their team's paddlers for stronger paddlers to stack their boats? Seemingly this is becoming a more prevalent item. If you've got part of your boat filled with Preds you're obviously going to fair quite better than with you're original team.

Don't post rankings with if you are going to half-ass the measurement criteria. It just cheapens the sport."

What an idiotic comment. I agree with the previous poster... if you don't like it, see if you can do better. But first you should make sure you understand the methodology before you describe it as "half-assed".

Regression is used to establish relative rankings. As such, the impact of factors such as weather and course differences are considered, since these factors should have a similar impact on all crews that race in an event. Clearly having more data points, and more overlap between teams will improve the rankings.

As discussed above, changes in personnel (particularly temporarily importing "star" paddlers) will skew the results. So what is your suggestion to get around this? Require that every team formally submit a roster on April 1st and allow no substitutions?

I think they've (including Paddle Demon) done a great job, and rather than cheapening the sport, it has added an interesting (an always thought provoking) element.

By the way, it is "your", not "you" ranking scheme, "fare" not fair, and you should learn when to use "you're" and "your". Sorry to be pedantic, but your negative and stupid post pissed me off.

 
 
Anonymous

Re: Rankings update

September 20 2009, 3:42 PM 

Perhaps you missed the message title. "Simply a joke"
Baited.
Caught.

 
 
Anonymous

Re: Rankings update

September 20 2009, 3:49 PM 

"Perhaps you missed the message title. "Simply a joke"
Baited.
Caught."

I don't think so... I think the title was reflecting the poster's view of the rankings. You're right... it could have been a reference to their education.

 
 
Anonymous

Re: Rankings update

September 20 2009, 4:51 PM 

Anon 3:42. Epic Fail. You are a moron.

 
 
Anonymous

Re: Rankings update

September 21 2009, 8:33 AM 

"Further, you ranking scheme is skewed. Have you included any weather components? Have you normalized for differences in the courses? Did you even consider teams who prefer to ditch their team's paddlers for stronger paddlers to stack their boats?"

Maybe read the methodology before talking smack.

Have you included any weather components? YES!

Have you normalized for differences in the courses? YES!

Did you even consider teams who prefer to ditch their team's paddlers for stronger paddlers to stack their boats?" YES - rankings based on speed so teams that can stack and go faster subsequently get ranked higher.

Next.

 
 
Anonymous

Re: Rankings update

September 21 2009, 12:33 PM 

Re: Rankings update September 20 2009, 4:51 PM

Anon 3:42. Epic Fail. You are a moron.

Hahahaha! this is too funny!

 
 
Anonymous

Re: Rankings update

September 21 2009, 1:12 PM 

I get it, someone wrote something that wasn't true.

Hahahaha. Good one!

 
 
Anonymous

Question to MJC...

September 21 2009, 3:59 PM 

Thanks to all those who gave their opinions. Right you may be, but I would still like to hear from him. So, in fear that there have been so many responses by now that my original post is "lost" in the mix, I will re-post my original question to MJC:

MJC, just wondering.

How do you decide which team is seeded higher when more than one team end up averaging the same time? For example, I notice that there are three teams at 2:11:3 (RBC Comets, Victoria College, and PDBC U23 I). I would have thought that Victoria College would be ranked above the other two simply because they have more regattas included in their rankings (the other two teams only have one regatta included in their rankings while Victoria College has three), making their average time more "reliable".

Is there another variable you are using, like having more decimal places than what you are posting? Or maybe there's another "tiebreaker" you are using (it doesn't appear to be alphabetical).

I used this only as an example, the same scenario plays out for other set of teams (see 2:04:2, 2:07:4, 2:07:9, 2:08:1, 2:09:1, 2:12:3, etc.). In all these cases there are teams that have more sample regattas seeded under a team with fewer.

Thanks.

 
 
Anonymous

Re: Rankings update

September 21 2009, 4:17 PM 

The better looking team gets priority.

 
 
Anonymous

Re: Rankings update

September 21 2009, 4:35 PM 

"How do you decide which team is seeded higher when more than one team end up averaging the same time? For example, I notice that there are three teams at 2:11:3 (RBC Comets, Victoria College, and PDBC U23 I)."

Does it matter? I really don't think it is even worth posting about if the 59th ranked team gets the 61st rank due to a tie. I can't imagine it is in the rankers interest to take an hour of programming to come up with some "special formula" for ties.

These rankings may be good but they are not accurate to 0.1 of a second anyways - maybe 1 second. So rolling a dice would be about as fair as anything else. I'd focusing on what your team needs to do in order to get 0.1 seconds faster, nobody has 2:11.2.

 
 
Anonymous

Re: Rankings update

September 21 2009, 5:48 PM 

Would it "matter" if the tie was for first? I was just wondering if it was random or if there actually was a science to it. If not, why not rank them: 59, 59, 59, 62 (for example) to reflect a tie.

Again, just wondering that's all.

 
 
Anonymous

Re: Rankings update

September 21 2009, 10:02 PM 

Great ranking job! I looked at it for the first time today. It seems to be very acurate. Why is the Lachine Knockout and QC Cup not included? With so many Quebec teams in the rankings it would only make sense. Both those festivals are heavily attended by Qc sport teams.

Keep up the good work!

 
 
MJC

Re: Rankings update

September 24 2009, 10:15 PM 

The Quebec Cup and Stratford have been added to the rankings

Anonymous 6:08pm, The rankings website shows times formatted to the tenth of a second but as somebody else correctly suggested there are more digits used in the model so in the event of a tie we use these less significant digits to determine the order.

Anonymous 11:02pm, thanks for pointing that out, we had both "Entre Amis" and "Entre Amies" for the women's crew. This has been consolidated under "Entre Amies".

Anonymous 12:44am, yes the rankings do normalize for course differences and (sustained) weather differences between regattas. Gusts of wind that benefit only some teams can skew the results slightly. This is one of the reasons why we want to include as many races as possible - so factors like wind get averaged out. If you race often enough you should get a fair share of head wind and tail wind. You can read the methodology section to understand how the results are adjusted.

Anonymous 10:02pm the Quebec cup is now added. The rankings only include 500m races so Lachine is not included. I put together a separate ranking that includes all race distances (similar to what Power Demon did) and that does include Lachine but I haven't had a chance to check it for errors and put it on the website. I still am not sure if we should go with 500m only or all race distances next year.


btw I took the liberty of renaming "Lloyd Halyk Memorial Dragons" to "Hanalei" for the rankings. I hope this is OK. Similarly, can somebody tell me if "MICHEL PELLERIN" is an existing dragon boat team (Perhaps 3R Mixed A?)

Thanks!




 
 
Anonymous

Re: Rankings update

September 24 2009, 11:08 PM 

Michel pellerin is 3R A

 
 
Anonymous

Stratford, where statistics go to drown in shallow water

September 25 2009, 9:49 AM 

While I personally don't really think Stratford should be included in the ranking due to the crazy lanes and conditions (Can't avoid riding or being ridden even if you wanted too), the differential between Hanalie and (Lane One, BMO and Blades) doesn't add up.

The actual Diff between Hanalie and Lane One was 3.92s and it's 6.8 in the rankings

The actual Diff between Hanalie and Blades was 5.53s and it's 9.4s in the rankings

The actual Diff between Hanalie and BMO was 6.17s and it's 10.4s in the rankings

If anything, since the times are adjusted downwards, you would think that the differentials would also shrink and not grow.

Maybe a good reason to exclude it entirely. I can understand that there may be some cross event weighting affecting, BMO and Blades, but since this is the only event Lane One does, I don't see how their time isn't a direct correlation to their differential to the others.

In any case, as I said, I would be just as happy if the event wasn't included, but as is, it may be identifying some kind of anomaly in the formula/algorithm.

 
 
Anonymous

Re: Rankings update

September 25 2009, 10:15 AM 

Agree with the above regarding Stratford. It should be excluded from the rankings but included in the archive for informational purposes, as the results do tend skew things up a bit.

 
 
Guru.

Re: Rankings update

September 25 2009, 10:22 AM 

How about all of you shut up and let MJC do whatever he wants. First people want it, then they don't want it. At the end of the day, the results shouldn't be changing placement very much if you've had a few races. Why do you people care so much? It's used as a proxy to see where your team compares to other teams. You don't get a trophy or money for being at the top of the list. The only thing that should be corrected is name matching. Let the model tell the rest.

 
 
Anonymous

Re: Rankings update

September 25 2009, 10:36 AM 

Not sure if you are referring to my feedback on the times or references to it being excluded, but as I said it really doesn't matter to me much.

So I agree as far as including/excluding that it is entirely up to them. I think we are lucky to have it period. I think they have done an awesome job taking this on and striving to make it the best it can be. I also agree that many take it way to seriously.

I can assure you though that from a times perspective, the rankings folks take a lot of pride in their work and if there is something a little weird they would want to know about it and take a look. From my perspective, if we have it, it's worth identifying things that seem a bit strange so they can look at it if they wish.

 
 
Anonymous

Re: Rankings update

September 25 2009, 10:41 AM 

"How about all of you shut up and let MJC do whatever he wants."

"Guru", MJC asked for feedback, and they're getting it, and pretty constructive feedback.

Can you offer something constructive too?

 
 
Anonymous

Re: Rankings update

September 25 2009, 11:34 AM 

Hey Guru are you trying to be a f***tard?

I never liked Stratford being added, and I'm not one of those that say "I told you so", but I told you so lol. Kidding aside, some people believe Stratford should have been added due to enough teams going to the event to do so. MJC agreed there was enough for it to be added. Now someone else is showing some Stratford results that appear to skew things, which might validate why it shouldn't be added. I'm not sure Power Demon ever had Stratford added in his true rankings, but if he didn't there could be interesting reasons.

MJC can do what he wants as he's the keeper of the new rankings, but that doesn't mean people can't provide input. No one was really complaining until you showed to tell people to shut up.

 
 
"Power Demon"

Re: Rankings update

September 25 2009, 2:50 PM 

"I'm not sure Power Demon ever had Stratford added in his true rankings, but if he didn't there could be interesting reasons."

Initially I included everything I could but later found that for many of the smaller festivals teams didn't show up with their full crew, or merged boats, problems with the course, and with a low sample size the results seemed to be more skewed than improved.

In 2008 I only started ranking festival that had a significant competativeness score (i.e. a certain % of top crews, weighted by rank). This not only saved time but made for better rankings IMO. Stratford got dropped in 2008 as a result.

Here is what was included after the competativeness rule for 2008:
Pickering, Lachine, Toronto Island, Ottawa, Nationals, Montreal, GWN, and Quebec Cup
Sudbury and Woodstock were on the verge.


I recommend the same, come up with some simple rule based on attendance on whether to include or exclude a festival. Otherwise you spend a lot of time on something that may not be improving things. A lot of work has already gone into producing a great ranking for 2009 so keep it simple and hopefully the rankings will stay alive for a many years more.

 
 
Guru.

Re: Rankings update

September 25 2009, 2:50 PM 

Bah, quit your whining.

 
 
MJC

Re: Rankings update

September 28 2009, 11:12 PM 

Thanks for weighing in PD. Filtering regattas based on competitiveness seems like the right thing to do for the rankings going forward. We have a competitiveness measure based on attendance already that we can use. By the way if you want to take over the rankings in the future fire us an email.


Anon 9:49AM, Stratford is a relatively shallow/slow race course so you are correct that times for all teams are adjusted downward but that does not mean the difference in times between teams needs to decrease after adjustment. Remember the adjustment used is a variable that depends on a team's speed. A fast team can be adjusted by more than a slow team if the data implies it. That is what is happening with Stratford. You can see the methodology section of the rankings website for more details. What you (seem to) have in mind would be an adjustment that is a straight horizontal line.

Imagine a large regatta at a deep, fair course like Welland or Banook. Now imagine the same regatta is held on a shallow course with tight lanes. A team that won handily on the deep course may win by only a fraction of a second on the shallow course. The best adjustment for the shallow course may actually increase the spread between teams as would have been the case for the deep course. (We don't label a course as shallow or deep, the adjustment is 100% determined by the data.)



One potential problem with including results like Stratford is that although there were almost 20 teams in common with the rankings prior to Stratford that are used to determine the adjustment for Stratford there were few very fast teams in common so the adjustment curve at the fast end is not determined by many data points. Ideally to add a regatta to the rankings we want to have lots of teams in common and distributed such that there are teams of various speeds in common. Using Power Demon's filter to include only competitive regattas in the rankings should avoid this potential problem.


 
 
Anonymous

Re: Rankings update

September 29 2009, 8:23 AM 

"...Stratford is a desperately shallow/slow race course..."

2:20's? I hurt just thinking about that.

 
 
Anonymous

Re: Rankings update

September 29 2009, 9:36 AM 

I would have thought that each event had a weighting / adjustment and not each team per event got a separate adjustment. Or at least that it would seem logical that each team would be roughly effected the same for each course and would result in relatively close adjustments.

All approx: Hanalie was adjusted downwards by 13.5s, Blades by 12.5s, BMO by 12s and Lane One by 10.5s. So from one extreme to the other there is an adjustment difference for the course of about 3s. Maybe Lane One is a bit of an outlier because there are no other events for them. If you take them out, it is a lot closer.

 
 
Anonymous

Re: Rankings update

September 29 2009, 9:51 AM 

"Or at least that it would seem logical that each team would be roughly effected the same for each course and would result in relatively close adjustments."

Not the case at all. Longer course doesn't hurt top fit crews nearly as much as weaker crews. Shallow water tends to pull crews together while deeper water spreads crews apart. You need to adjust the different competative tiers differently.

 
 
Anonymous

Re: Rankings update

September 30 2009, 6:26 PM 

A pro-Stratford opinion...

I feel strongly that ALL 500m regattas should contribute to the rankings. My reasoning is simple: different teams attend different regattas, so excluding regattas could exclude important results for some teams.

Furthermore, there is a judgement call involved as soon as you start to exclude races. Basing exclusion on how many high-end teams de-values the lower teams (and there are teams that are ranked lower than 50 or even lower than 100 that are still very interested in the rankings).

As a compromise, I like the idea of weighting regattas to determine a team's average. The current method of ranking based on "competitiveness" is a good start, and I much prefer it to endless arguments about which regattas should or should not be included. Another possibility would be to weight regattas based on "fairness" so regattas with heavy lane imbalances (Pickering) or with iffy timing (lively events) could be weighted as less - note that the data to support such a weighting system would have to be built up over years of course analyses, so sticking with the current system seems prudent.

Maybe a solution to the time-adjustment-by-speed problem is to combine results over the next few years of regattas to determine the adjustment line. This requires the assumption that each year's teams are independent of their previous teams, which isn't that bad of an assumption... It would allow for more data points in all speed ranges, which could help to make future years' rankings more accurate.


Again, please don't remove regattas from the rankings - that opens another huge can 'o worms.

 
 
Current Topic - Rankings update  Respond to this message   
  << Previous Topic | Next Topic >>Return to Index  
Find more forums on Water SportsCreate your own forum at Network54
 Copyright © 1999-2014 Network54. All rights reserved.   Terms of Use   Privacy Statement