New Posts New Posts RSS Feed: Measurement Based Handicapping
  FAQ FAQ  Forum Search   Register Register  Login Login

Measurement Based Handicapping
 Post Reply Post Reply Page  123 7>
Author
  Share Topic Share Topic  Topic Search Topic Search  Topic Options Topic Options
Isis View Drop Down
Really should get out more
Really should get out more
Avatar

Joined: 01 Sep 05
Online Status: Offline
Posts: 2753 		Post Options Post Options   Quote Isis Quote  Post ReplyReply Direct Link To This Post Topic: Measurement Based Handicapping
    Posted: 25 Nov 08 at 5:38pm
Hi folks,

Apologies if this is a bit long and rambly - it's been a long day.
I'm just getting started on my dissertation at the moment, looking at the performance prediction of dinghies and the feasibility of a new measurement based handicapping/rating system.

The commonly used PY system is a single number based on club returns and lots of number crunching to predict performance based on past results.

The aim of this project is to produce a framework for predicting the performance of various boats based on a limited number of key dimensions. Think IMS instead of PHRF if you are a big boat sailor.

I'm not expecting to produce a ready-to-go system in the timeframe I have but rather to look at how feasible it would be, and to assess the pros and cons of such a system vs PY and its variants around the world.

For the time being at least this will not deal with multis/foilers but the aim is that it will be able to model everything else from Fireflies to 49ers with some degree of accuracy - hopefully with the more traditional designs being modelled quite closely.



So, no formal questionnaire but it would be great to hear a few opinions on the subject, particularly after the 'handicap bandit' threads...

Handicapping is never perfect and will always favour certain designs over others but with that in mind: how happy are you with PY handicapping? Just a bit of fun, unfair, or great?

What do you think about single number handicaps? If it were possible would you rather different numbers for varying wind strengths/courses? Would your club deal with that extra hassle?

What about individual boat handicaps to take into account variations within classes (crew weight?) or again, too much hassle and too hard to enforce?

If the project shows that a reasonable handicapping system is possible - Should that system be open to designers to actively optimise for a better rating or should it remain a 'black box' rule?



As the project progresses I will be contacting class associations for help with measurement details but if you would like to help by spending half an hour with a tape measure one Sunday morning - or even better if you can offer some gps data with some weather information then please get in touch. Similarly if you represent a class association and would like to know more then let me know!
Back to Top
Chris Noble View Drop Down
Really should get out more
Really should get out more
Avatar

Joined: 26 Nov 04
Location: Scotland
Online Status: Offline
Posts: 710 		Post Options Post Options   Quote Chris Noble Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:00pm
oh dear ben... its definetly time for you to hit the sailing team social tonight and hit it hard
Competitive Boat Insurance From Noble Marine

FOR SALE:

I14 2 Masts 2 poles 3 Booms, Foils Kites/Mains/Jibs too many to list.
Back to Top
JimC View Drop Down
Really should get out more
Really should get out more
Avatar

Joined: 17 May 04
Location: United Kingdom
Online Status: Offline
Posts: 6662 		Post Options Post Options   Quote JimC Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:08pm
One hell of a project... I'm not sure its feasible.

I reckon PY works a lot better than most people think... As is probably obvious! When you do the numbers you usually find that changes that would make much difference are ridiculously large...

For my money one number handicaps are essential, because I can't think of any way you can have multi number handicaps without introducing a human element. I'd rather each dog have its day than a) trust the race officer to get the evaluation of whether its number series a, b or c to be right or b) have to be the race officer who decides which number set to use...
Back to Top
bert View Drop Down
Really should get out more
Really should get out more
Avatar

Joined: 23 Apr 05
Location: norwich usually
Online Status: Offline
Posts: 584 		Post Options Post Options   Quote bert Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:13pm

Sounds like a hell of a job -- good luck

So, no formal questionnaire but it would be great to hear a few opinions on the subject, particularly after the 'handicap bandit' threads...

Handicapping is never perfect and will always favour certain designs over others but with that in mind: how happy are you with PY handicapping? Just a bit of fun, unfair, or great? - pretty good considing how many clubs do send in returns.


What do you think about single number handicaps? If it were possible would you rather different numbers for varying wind strengths/courses? Would your club deal with that extra hassle? -- It would produce more even results with the conditions of the day to be taken in to account for each class,However the OD would have a fit & I can`t see that happening at more then 1% of cases,

What about individual boat handicaps to take into account variations within classes (crew weight?) or again, too much hassle and too hard to enforce? -- as above,The existing system allows for this via personal handcaps to be applied by the clubs.

If the project shows that a reasonable handicapping system is possible - Should that system be open to designers to actively optimise for a better rating or should it remain a 'black box' rule? -- if by "black box" you mean - this is the system,do not alter it then I would ask why use it instead of the existing system which already allows "local" adjustments,So the answer in my veiw is YES the system should be allowed to be open to designers but the existing system already allows " designers" to adjust for local considions or personal handcaps already.

 

Sorry if that sounds negitive but I wish you the best of times with it all.
Phantom 1181
AC-227 IC 304
blaze / halo 586


Back to Top
Guest View Drop Down
Newbie
Newbie


Joined: 21 May 04
Location: United Kingdom
Online Status: Offline
Posts: 0 		Post Options Post Options   Quote Guest Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:15pm
Just sounds like another way to reach the conclusion that class racing is what it's all about and anything else is just for laughs
Back to Top
Guests View Drop Down
Guest Group
Guest Group
Post Options Post Options   Quote Guests Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:22pm
Isis, Pm Medway Maniac. I believe he's created a PY predictor
formula and I'm sure he'll be happy to share his findings. It might
help a bit with your dissertation....?

Edited by Doug.H
Back to Top
alstorer View Drop Down
Really should get out more
Really should get out more
Avatar

Joined: 02 Aug 07
Location: Cambridge
Online Status: Offline
Posts: 2899 		Post Options Post Options   Quote alstorer Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:35pm
Originally posted by Guest#260

Just sounds like another way to reach the conclusion that class racing is what it's all about and anything else is just for laughs


I'd agree with this.

How about input from the cat and yacht sailors? I'd reckon that next to PY that IRC is possibly the other "main" handicaping system in the UK- a "closed" measurment rule, that is often updated in order to limit loopholes, and that promotes a certain sort of boat (ones that can actually be used for crusing too, rather than all-out stripped out racers). Then there's the cats which have the SCHRS, an open rule rating system that you might want to look at if you're trying to come up with something for mono hulls.

Also, don't forget to consider PY systems in other countries such as the US and Aus.
Back to Top
Isis View Drop Down
Really should get out more
Really should get out more
Avatar

Joined: 01 Sep 05
Online Status: Offline
Posts: 2753 		Post Options Post Options   Quote Isis Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 6:38pm
Thanks for the responces already - I will try and get back to those of you who have PM'd me and reply in more detail tommorrow as the aforementioned sailing social is looming and im going to be in trouble if im late....



Just to clarify though by 'Black box' I mean that the workings of the rule are hidden to the public eye.

IRC is a good example of a Black Box rule - Your measurements go in and your rating comes back out with little detail of how the number was arived at. This means that it is largely unknown which factors will affect your rating most.

IMS on the other hand is an open rule - Designers have access to the VPP (velocity prediction program) which produces the number and can therefore focus their efforts specificaly towards that rule.
Back to Top
Daniel Holman View Drop Down
Really should get out more
Really should get out more


Joined: 17 Nov 08
Online Status: Offline
Posts: 997 		Post Options Post Options   Quote Daniel Holman Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 7:18pm
Mate PM me I put together some spreadsheets a while ago that could be of use / interest to you. Basically took commonly available parameters and assumed some crew size / weight parameters ect, found the approx sensitivity of each parameter by regression, and used these regressions to have a pop at a formula for predicting PY for the given parameters. Slightly different formulae for bots with and without kites. The formulas can give a better fit if narrowed down to each boat type, e.g 2 person hiking no spin. I just did them for my own interest to get a handl on the sensitivity of a boats speed round a course from tweaking the basic parameters. Again it cannot be perfect as it cannot account for how highly tuned ect the boat is. Also, one of the main sources of "banditry" in my method is WL beam - basically as the vast majority of dinghies push water most of the time, ones with a narrow WL do better than their other parameters would suggest i.e moths (non foiling) canoes, RS300s ect, so unless there could be a reliable way of ascertaining dinghy WL beam data then that would be an easy rule cheating ploy, black box or not.
As has been previously mentioned, PY as it is is pretty good for widely sailed classes at "average" venues in non extreme conditions, but obviously getting away from this and the banditry gets a little rife!
Back to Top
Scooby_simon View Drop Down
Really should get out more
Really should get out more


Joined: 02 Apr 04
Location: United Kingdom
Online Status: Offline
Posts: 2415 		Post Options Post Options   Quote Scooby_simon Quote  Post ReplyReply Direct Link To This Post Posted: 25 Nov 08 at 8:04pm

Originally posted by Isis

Hi folks,


For the time being at least this will not deal with multis/foilers but the aim is that it will be able to model everything else from Fireflies to 49ers with some degree of accuracy - hopefully with the more traditional designs being modelled quite closely.



You mentioned measurement based systems.......I might know a little about one...

We have 2 for Cats

1, SCHRS which is also trialling the Weta Tri

2, Texel.

 

I've actually done a fair bit of thinking about how SCHRS might model Mono's.  Far too much to write down here, but feel free to PM me your Skype or Phone number and I'll give you a call.

 
Wanna learn to Ski - PM me..
Back to Top
 Post Reply Post Reply Page  123 7>

Forum Jump Forum Permissions View Drop Down

Bulletin Board Software by Web Wiz Forums® version 9.665y
Copyright ©2001-2010 Web Wiz
Change your personal settings, or read our privacy policy