Well, yes, something like that.

Or the fact that I choose to win in another field over winning in the F16 class !

Sometimes people have to make choices and I just made mine. And the other option is a whole lot more attracttive financially.

Wouter

That's a shame, I was looking forward to meeting you... When I had a real job I did some research on data fusion in a vision (robotics) context using Kalman Filtering as the root technique - and I can't help feeling that it could be useful in a rating context; a possible way to fuse model-based and empirical techniques. We had great success with it in vision; the only problem is that I just looked through some old stuff and it looked like gobbledegook...!

So was I.

Come on Wouter, come over and Play!

I know the "Kalman filter" mostly in its application as an "observer" in control systems. However, its use there is indeed to estimate the values of certain states that can not be directly measured. It is as such an estimation technique.

After thinking about it some more, it does strike me as a parameter estimation (weighting coefficients) procedure after the designer has chosen the model structure. As such it falls in the very same (regression) approach as Texel and SCHRS deploy, although they use a different regression technique.


I just looked it up in Wikipedia and it reads like something I should understand. Sadly I think some pages in Wikipedia have been hijacked by mathematic nerds and rewritten in their own special secret language.

Leave it to formally trained mathematicians to write down simple stuff in the most confusing way.

Anyway, the general approach you are suggesting is in my opinion the best way to build a rating system. Choose a model structure of a relative low order and then use an estimation technique (like Kalman filtering) to estimate the weights. Of course all ratings will then have to be calculated in exactly the same way as Texel and SCHRS do, by measuring boats and punching in the specs in the regressed formula.

This is significantly different from how some posters understand this intention in sailwave.

Summerized it is what Texel and SCHRS are doing only you (Sailwave) is proposing to use an estimation technique that continiously updates itself with each new data coming in.

It that is the case then you still have the problems associated with undependable datasets. What if I enter the results of my last (drunken) sunday fun club race ? It appears to me that Kalman filtering will respond to that when in reality it should not.

Interesting problem. But there are always next years !

Wouter

Scooby Simon when are you hoping to update the site? sailing season if fast approaching. Did anybody read the latest from the RYA on their new PY web site and update system.
Taken from the Daily Sail:


Many of those racing dinghies at various clubs around the world have come to accept that handicap racing is far from perfect. However, the RYA, who are charged with managing the Portsmouth Yardstick handicap system in the UK are now looking at wholesale change that should improve this throughout the country. To understand what changes are being made it is important to first understand how the previous system worked. Principally it relied on sailing clubs returning the results from their racing every year, by post to the RYA. These results would then be handed to analysts who would look over all the data and, using specific formulae, adjust class handicaps accordingly. Theoretically this system works well; however, it relies on all sailing clubs storing their race data for the season and sending it in one go to the RYA. As one might imagine not many clubs made their returns and the less data there is the less accurate the system is. Such a broad ranging system also allows for no regional variations and as most will appreciate sailing an International 14 on a small lake in the midlands is very different to sailing one in Weymouth harbour, for example.

In recent years the PY system has started to be even less relevant again to the dinghy sailing community. Principally this hinges on the speed at which the handicap system can be updated. If one considers the case of the International Moth, there is simply not enough information at the RYA to give a meaningful handicap to the class. However, boats have been racing on foils in the country since 2004 and most would argue there must be data somewhere to conjure up a relevant handicap. Also with development classes such as the Moth they can change so quickly a yearly returns system realistically cannot give a reliable up to date handicap for the class.

In addition to this, given that modern, high performance dinghies can be sailed at such different speeds, there are those who argument as to whether one class handicap can be fair at all. In February, for example, we saw Graham Vials win a race at the Tiger Trophy where he lapped all but a few of the rest of the Moth fleet. Can a single class handicap really be suitable with such a spread of speed throughout a fleet? Finally, though on a similar subject there has been much debate as to whether classes can and should have various handicaps in various wind ranges. Continuing with the Moth example, a foiler in five knots not up on the foils is going to be very slow, whereas in perfect 15 knot conditions it will be very fast. The question is then should there be various handicaps to reflect performance in various wind ranges?

There are a great deal of other issues in designing a handicap system and it is fair to say there is no one, single answer that will make handicap racing perfect. Happily, though, the RYA are now trialling a new system that might well go some way to remedying some of these issues. Charged with this mission is the recently appointed Technical Officer for Racing, Sebastian Edmonds..

Given that clubs are forwarding less and less race results data to them Edmonds explains that core to the new system is making it easier for clubs to transmit this information back to the RYA. “Essentially we have come up with an RYA results trial site, where clubs can submit the results of their racing online, instantly,” he says.

The site Edmonds refers to is already up and running, though in a trial stage. It is compatible with most types of spreadsheets as well as HAL and SailWave results programmes and is being developed in conjunction with SailRacer enabling an online database of events and results. In theory clubs should just be able to upload their results instantly and in the format they usually use to work out and publish their results. This makes the delivery of data to the RYA significantly easier and should mean much more data gathered which, in turn, means more accurate handicaps.

The new system does not just make returning information to the RYA easier though; there are many other benefits and potential benefits that come with it. “When a club submits their results to the system they get an instant suggested handicap for each class,” Edmonds explains. “This means the RYA will be able to offer ‘Club Yardstick Lists’ based on the sailing conditions experienced at that club. A good example is the performance difference of an Optimist or Thames A-Rater sailing on an inland lake or at Brighton – both would perform very differently on either sailing waters.” Edwards adds that it would take something in the region of 20 races before a new individual ‘club yardstick’ is accurate.

This is clearly a massive step forward, enabling all clubs to receive their own class specific handicaps that relate directly to the fleets sailing at that club. Clearly then, this will make handicap racing a more accurate and so more pleasing form of racing. The benefits to clubs do not stop there either, as the results can then be easily posted back onto a club’s own website from the online database.

It is fantastic to see the RYA making such a significant change to the handicap system and one that would seem to benefit everyone. It is even more encouraging that this system in not just being talked about but is actively up and running now on a trial basis with the plan to roll it out nationwide in 2009. However, as much as this is good news in the short-term, there are longer term goals that are more exciting still.

There are a number of different ways the new system could be used in the future. “To start with we are just looking at getting more data from clubs with this new system,” Edwards states. “In 2009, though, we will be trialing individual or personalised handicaps through the same software.” He adds that creating a database that handicaps people individually - ie by their name - is difficult, but doing so for a sail number in a class is very doable. Given most sailors sail with the same boat week in week out personal handicaps could represent a fantastic step forward in the way handicap events work.

Further to this, with such a system in place would it not be possible for clubs to submit wind strengths for races and so give handicaps based on the conditions racing takes place in? “It would depend on the competence of the club, but it is certainly possible to add wind data to the system.For the moment though we are just focusing on making data gathering and handicapping a simpler, more effective process,” Edwards comments.

Certainly once wind strength is include the system will get a lot more complicated as there would, presumably, have to be a decision made about what wind strength band the racing was taking place in. Also most races in the UK does not take place in a single, specific, wind strength, further complicating the matter. It may be that this is one step too far for many and certainly there does seem to be a little hesitation from Edwards when discussing these, rather more complex ideas. He is keen to express that though this and the personalised handicaps are possible, for the moment it is important to get enough data back from 2008 to ensure a decent set of handicaps can be delivered by the end of the year, with personalized handicaps being trialed in 2009.

Of course all this new technology is great and the system could well deliver one of the most complete and efficient handicapping systems we have seen. However, none of this changes one simple, fundamental fact: The accuracy of a handicap depends entirely on the volume of data used to work it out. In short unless clubs use this system to submit their results, in spite of all the bells and whistles and possible uses of the new site, we will not have any more accurate handicaps than we use now. The one major difference here is that by providing clubs the option of having their own handicaps, based on their specific racing, the RYA are offering an instant carrot for clubs all over the country. The big question now is, is the carrot big enough to be worth the, reduced, bother of submitting returns?

We have a minor tweak to the SCHRS system that will be introduced as soon as myself and members of the team can get the appropriate data together and update the website.

Very few numbers will change.

Hopefully updates will be done soon(I cannot commit to a date at present) - the day job is getting in the way of everything at preset. Blame the bloke who worked for SocGen!

Trust me, it's not earth shattering, don't hold your breath!