RYA National Handicap for Cruisers.

[DJE]

I'm seriously thinking about implementing this for our cruiser fleet this season. The first problem is finding base numbers. Many of our boats are not on the published list but we have PY numbers that we have been using and adjusting for several years so I would like to convert them to NHC numbers. Foolishly thinking that this would be simple I downloaded the current NHC list and the 2012 PY list from the RYA and compared the two. All the boats that appear on both lists are plotted below; the scatter is amazing. The obvious conversion I thought would be NHC = 1000/PY. That is the yellow line and it is a long way from most of the data points. The other lines are my attempts to fit a formula through the points. Anybody else struggling with this?

 

[AntarcticPilot]

I'm seriously thinking about implementing this for our cruiser fleet this season. The first problem is finding base numbers. Many of our boats are not on the published list but we have PY numbers that we have been using and adjusting for several years so I would like to convert them to NHC numbers. Foolishly thinking that this would be simple I downloaded the current NHC list and the 2012 PY list from the RYA and compared the two. All the boats that appear on both lists are plotted below; the scatter is amazing. The obvious conversion I thought would be NHC = 1000/PY. That is the yellow line and it is a long way from most of the data points. The other lines are my attempts to fit a formula through the points. Anybody else struggling with this?

Looking at it purely as a matter of data analysis, I'd say that the residuals of any line fitted to those data will have too great a scatter for the exercise to be worth doing. Did you compute an R-squared value? If you look at a restricted interval (e.g. between 1000 and 1100 PY), the scatter in the corresponding NHC values is about 0.15 - and the total range of NHC values is only 0.3! So, the errors from any curve-fitting exercise are going to be about 50% of the possible range of values.

In other words, there is effectively no relationship between PY and NHC - which is crazy! But then again, I think racing is crazy

Isn't one of the "features" of the NHC that it changes according to race results on a boat by boat basis?

 

[DJE]

Thanks for the comments. Yes NHC adjusts the handicap numbers after each race to reflect the relative performance of the boats. It saves me from having to decide my own adjustments at the end of each season (and allows me to hide behind the RYA).

My current plan is to use one of the fit lines to get base numbers and then run the adjustments through the last couple of seasons' results to produce starting numbers to use this season.

 

[Plevier]

I have found Bas Edmonds at RYA pretty quick at coming up with numbers for unlisted boats (although I'm waiting for one at the moment) on email request.

 

[Greenheart]

Any chance you'll name the classes concerned, DJE? As a current non-racer of dinghies, I'm nevertheless fascinated by the reliance on PY figures by huge international fleets, and the reputed inaccuracies they support, and the much greater ranges of performance which cruisers of any particular type must exhibit, given that their principle use isn't competitive and must include innumerable non-standard add-ons which blur the relevance of all tables of race results...

...quite possibly my interest here is likewise irrelevant, but standardising classes into performance groups is a preoccupation of mine. I guess you're familiar with Byron cruiser tables?

View here: http://www.byronsoftware.org.uk/bycn/byboat.htm

 

[Kelpie]

I've spent more hours than I care to contemplate mulling over handicap figures from PN to Byron to CYCA to owners' associations.
There is never a direct correlation. It's highly frustrating, and when calculating handicaps lands on you you have to defend your decisions. I always take the tack that I should 'show my working' but really you can find a favourable or unfavourable way of viewing almost any boat.

I was hoping that the new RYA system would be a improvement, but I guess not...

 

[Piddy]

Hi DJE,

We did do exactly this at a Club not far from you. There was a general level of dissatisfaction of the base list with the boats designed for ratings rules (whether IOR, IRC or Ianything else) coming out favourably against those built as cruisers, so we converted our PY to get base numbers and started all series from that.
Of course the numbers aren't portable between clubs but none of our regular racers want to do that anyway.

Bas Edmonds was surprised when I bought that up at his presentation at our club, believing they had erred the other way. However the changes for this year, ironing out the extreme result and changing some of the calculations used to get the 'new' handicap (for the next race) will help.

I believe it has helped our racing - whilst those that had generous PY numbers are peeved that they are now judged on results, we have had more winners, a real closing up of corrected times and still the better sailed boats do win.

 

[DJE]

Any chance you'll name the classes concerned, DJE?

We have a cruiser fleet that ranges from Newbridge Navigators to a Dragonfly 30 trimaran. Neither of those is on the RYA base list. Others missing are Beneteau 211 Beneteau 245, Legend 36 & Huzar 21. Pretty strange bunch really!


TBH it is the very wide range of boats and the strong tidal streams that we sail in which makes it impossible for any handicap system to cope adequately. I expect the ratings to jump about all over the place as one race might go up tide in light winds and the next one down tide in strong winds.

 

[RAI]

It always seemed odd to me to base vessel handicaps on race results where the skipper+crew combinations produce such different results even when sailing a single class. No surprise there is such a spread in NHC/PY numbers. Surely, to get comparative results one would need speed trials on different headings in different winds with different crews just to get polar diagrams of boat speed.

 

[Greenheart]

Presumably the virtue of the Portsmouth ratings is in the large number of race committees which report their results - so that regardless of untypical extraordinary results, the average positions of average boats in each class relative to another class, will stand up over time. Hence the fewer the results that are returned for less common classes (because of the rarity of these boats competing), the less reliable the handicap, because its position hasn't often been reported for comparison.

I'm not sure why I just said all that. :grey:

Interesting that boats can't be as easily categorised for comparative performance as cars - because variation in conditions so significantly affects which boat is favoured.

Although - given limitless modern number-crunching, perhaps the PY format could evolve to take into account minute-by-minute changes of wind-strength and sea state in the race zone, and the influence of each competing boat's excess displacement, extra propeller blades, sail age, until finally every boat's crew ability is accurately gauged, rather than the unsatisfactorily simplistic and invidious system presently used to rate the boats. I'm so glad I don't race.

 

[BlueSkyNick]

Dave,

we have gone for it hook line and sinker on purpose, whereas our neighbour club is persisting with the local handicap scheme.

My logic has been that I would rather have a system that works to some form of framework, albeit with vaguaries and errors, than a local system where one or two people are arbitrarily changing the ratings..... and still not getting it right.

It will take time for the NHC to be fully effective, and it works best over a series where all the same boats compete in every race, but you have to start somewhere.

 

[yoda]

Try 900/PY for the new base number. That's what our single handed lot in Plymouth are starting with and then using the system properly from there.

Yoda

 

[lpdsn]

They've probably used a different dartboard from the one used for IRC.

 

[Birdseye]

I typed out a long answer to the OP only for my tablet to say that my token had expired. I didnt know I had one!

Anyway, Mr OP, you cannot expect any correlation between your PY numbers developed after racing at your club with NHC numbers calculated from boat dimensions. For a start, the PY numbers will reflect crew competence so in the case of a regular race boat like a sigma with a decent crew the numbers will be far tougher than they would be for a similar boat raced rarely and with potterers driving it. This is true both club wise and nationally since some boats are common round the country and raced frequently whilst others are rarely raced.

In contrast NHC numbers are derived from a simplified version of the |IRC calculation and are the starting point fopr a performance handical system and not the end point like your PY numbers. Not only that but the system they have chosen is a bit crude and unlike PY does not recognise things like prop choice, twin vs fin keel, spinny or no spinny. Inmstead the NHC system is designed to adjust to actual performance very quickly which is not to say that boats like twin keelers starting off on the first race of a series wont be at a disadvantage to a sigma.

You have two practical ways forward. The first is simply to apply the NHC and let it adjust to performance. It does work or at least we have found it does. The second is to do a bit of simple maths. Select a fleet boat that is well sailed and is a half decent race boat - say something like a sigma 33. Calculate the reciprocal of its PY number and then work out a fiddle factor that will alter thae reciprocal to match the NHC for that boat. Then do the same for all your other PY numbers using the same fiddle factor. In effect you will have created your own list of NHC numbers as they would be after the last series.

Whatever you do your members will complain. If they are like ours you will find that every single boat is put at a disadvantage relative to the fleet under the new system. You have the british education system to than for that response.

P.S. There is a formula for converting IRC to PY and vice versa. ITs on the windows partition and I'm using linux. If windows still works, Ill find the formula and let you have it. It might be better than the simple reciprocal.

 

[DJE]

Birdseye,

Thanks for the reply. I would be interested to see your formula.

For the record the two sets of figures used to prepare the scatter graph were both published by the RYA. They were not figures for our club boats. Granted many of the numbers in the PY list were noted as "very low confidence" and based on very few returns. I have assigned base numbers to all the boats in our fleet (either from the published list or from the PY and the No.2 fit line on the graph) and run the corrections through the last two seasons of elapsed times in our club races - 12 races in all. I propose to use the ratings arrising at the end of this process as the starting NHC numbers for this season. If anybody doesn't like their rating they only have to turn up, do a few races, and watch it change!