Resetting the Fleets Median / Ratings

Imported from Yahoo
1

Hello,

Forgive me if this topic has been covered already but I have
a question regarding resetting handicaps when boats have drifted far above and
below the standard base line.

We have a situation where we have boats sailing 10% above
their NHC Base rating and also in the region of 10% below which can give new
entrants a huge advantage. Is it possible to reset the fleets ratings in order
to bring down the high rated boats and in turn bring up the low rated boats?

We have seen big changes in the handicaps from race to race, what % adjustment
should we expect and is it possible to set a limit to how high a boat can go?

Thanks,

Brian McElligott,

Iniscealtra Sailing Club

2

Hi Brian,

As far as Sailwave is concerned it will let you do anything you like with the Handicaps. You can specify the base ratings of your choice and also the rating used for the first race of the series for each boat. It is just a question of your committee agreeing the rules to which you are sailing.

When you say big changes could you send me a file (.blw) with an example and I can see if I can explain the change you are seeing. The calculations are done according to the RYA document. Sailwave does have the External/Debug mode which allows you to see all the calculation steps. In External mode you can do anything you like for the calculations so limiting it would be possible. It is also possible to modify some of the parameters that the internal mode uses for non standard use but lets have a look at an example first.

Jon

···

On 19 August 2016 at 09:12, bmcelligott1@gmail.com [sailwave] sailwave@yahoogroups.com wrote:

Hello,

Forgive me if this topic has been covered already but I have
a question regarding resetting handicaps when boats have drifted far above and
below the standard base line.

We have a situation where we have boats sailing 10% above
their NHC Base rating and also in the region of 10% below which can give new
entrants a huge advantage. Is it possible to reset the fleets ratings in order
to bring down the high rated boats and in turn bring up the low rated boats?

We have seen big changes in the handicaps from race to race, what % adjustment
should we expect and is it possible to set a limit to how high a boat can go?

Thanks,

Brian McElligott,

Iniscealtra Sailing Club

Jon Eskdale

03333 443377

07530 112233

3

Locally (in the USA), we use a system different from the NHC system but
our adjusted ratings often vary by 15-20% from the base - by which I
mean that some boats are 15%+ above base and others are 15%+ below base
(when converting to Portsmouth type ratings). The resulting differential
from top to bottom is essentially 30-40% of boatspeed

So, I'd say your plus/minus 10% is very normal for this type of system.
Of course, as a policy you can decide you don't want that. However, this
shouldn't be surprise as a well-sailed boat can easily be 1-1.5 kts
faster than a poorly-sailed boat. 4.5 kts vs 6 knots is a 25-33%
differential so plus/minus 10-15% should be considered perfectly normal
since the goal is to give every boat a chance to win NO MATTER HOW
POORLY SAILED.

If you impose limits I think you'll find that some boats never win a
race, which is inconsistent with a basic premise of this type of "golf
handicapping" system - which is to let every boat get an equal number of
1st-place trophies. Good sailors sometimes don't like this type of
system because it penalizes them for putting effort into becoming better
sailors. If I had to, I'd guess that if you are getting complaints it is
probably from the better sailors, who think they should win more often
than the truly horrible sailors. But, that isn't the goal or purpose of
this type of system.

Art

···

On 8/19/2016 1:12 AM, bmcelligott1@gmail.com [sailwave] wrote:

Hello,

Forgive me if this topic has been covered already but I have a
question regarding resetting handicaps when boats have drifted far
above and below the standard base line.

We have a situation where we have boats sailing 10% above their NHC
Base rating and also in the region of 10% below which can give new
entrants a huge advantage. Is it possible to reset the fleets ratings
in order to bring down the high rated boats and in turn bring up the
low rated boats?

We have seen big changes in the handicaps from race to race, what %
adjustment should we expect and is it possible to set a limit to how
high a boat can go?

Thanks,

Brian McElligott,

Iniscealtra Sailing Club

4

Hi Folks,

I have been doing back calculating the results and correcting them so the mean of the handicap = the mean of the yardstick, so. The keeps the ratings close to the middle of the fleet.

I see that Sailwave also uses a similar system to stop the mean fleet rating from drifting.

On a race by race basis the fleet leaders are more than 10% above the mean and I often start new sailors at -15% and adjust them up or down as I build up a data base. I have been having this experience from the mid 70’s

I know the NHC system talks about boats and this may be true for IRC boats but I and dealing with dinghies that have a nationally recognised yardstick, so it is obvious that the difference is the sailor rather than the boat.

I see the NHC system a good way of calculating personal handicaps only much easier for me and more formal. I publish both yardstick results and handicap results. The good sailors mostly consider the yardstick results and the other sailors consider the handicap results.

From a long series point of view the good sailors usually have a far higher attendance rate which tends to dominate a series results.

In our club we give both yardstick and handicap trophies.

I agree with Art that about 30% difference from the best to the worst at a club level can be expected.

Regards JohnL.

···

From: sailwave@yahoogroups.com [mailto:sailwave@yahoogroups.com]
Sent: Friday, August 19, 2016 8:31 PM
To: sailwave@yahoogroups.com
Subject: Re: [sailwave] Resetting the Fleets Median / Ratings

Locally (in the USA), we use a system different from the NHC system but
our adjusted ratings often vary by 15-20% from the base - by which I
mean that some boats are 15%+ above base and others are 15%+ below base
(when converting to Portsmouth type ratings). The resulting differential
from top to bottom is essentially 30-40% of boatspeed

So, I’d say your plus/minus 10% is very normal for this type of system.
Of course, as a policy you can decide you don’t want that. However, this
shouldn’t be surprise as a well-sailed boat can easily be 1-1.5 kts
faster than a poorly-sailed boat. 4.5 kts vs 6 knots is a 25-33%
differential so plus/minus 10-15% should be considered perfectly normal
since the goal is to give every boat a chance to win NO MATTER HOW
POORLY SAILED.

If you impose limits I think you’ll find that some boats never win a
race, which is inconsistent with a basic premise of this type of “golf
handicapping” system - which is to let every boat get an equal number of
1st-place trophies. Good sailors sometimes don’t like this type of
system because it penalizes them for putting effort into becoming better
sailors. If I had to, I’d guess that if you are getting complaints it is
probably from the better sailors, who think they should win more often
than the truly horrible sailors. But, that isn’t the goal or purpose of
this type of system.

Art

On 8/19/2016 1:12 AM, bmcelligott1@gmail.com [sailwave] wrote:

Hello,

Forgive me if this topic has been covered already but I have a
question regarding resetting handicaps when boats have drifted far
above and below the standard base line.

We have a situation where we have boats sailing 10% above their NHC
Base rating and also in the region of 10% below which can give new
entrants a huge advantage. Is it possible to reset the fleets ratings
in order to bring down the high rated boats and in turn bring up the
low rated boats?

We have seen big changes in the handicaps from race to race, what %
adjustment should we expect and is it possible to set a limit to how
high a boat can go?

Thanks,

Brian McElligott,

Iniscealtra Sailing Club

5

I agree that about a 25% spread in finish times seems to be normal for large one design fleets. The spread of finish times in Olympic races can be quite enlightening.

I haven’t studied NHC algorithms in detail (although I may have to) but when I ran tests against past club results it still seemed to be essentially “best sailor wins the series” rather than a true random scattering of prizes round the fleet. Certainly when I looked at using NHC as an alternative to our club personal handicap series it seemed to give the prizes to much the same people as the standard handicap, whereas our system (which we’ve retained) is basically a “most improvement on previous performance wins” system and produces different winners, which is what we want for dual scored series. Presumably there will be a lot of “Best performance on the day compared to usual level” about it.

Resetting NHC ratings seems a bit odd, since the base NHC ratings are very crude and of very limited accuracy. The fundamental reason for NHC was that the number of returns for cruiser Portsmouth Yardstick had diminished so much that the numbers were getting meaningless, and the system had lost the critical mass required for a statistically based empirical handicap to work. As IRC already provides a measurement based empirical handicap, and aims for as little measurement as possible anyway, my understanding was that after a good look at alternate systems round the world, and especially being inspired by Ireland’s Echo, the RYA came up with NHC as being the most viable option for entry level racing.

The alternative would seem to be people in smoke filled rooms deciding what handicaps ought to be, which seems like a recipe for the sort of allegations of cliques, corruption and cronyism you see on forums where they talk about the US PHRF system. I think it could be argued that f base NHC ratings were accurate enough that resetting to them regularly was a good idea then we wouldn’t need NHC in the first place, we could just use the base ratings.

6

Thanks for your input folks, it seems that in most of you seem to mange the system with very little manual intervention.

The only way of adjusting the median is by exporting the ratings and manually adjusting them and imputing them again? This to jimc_hjones point is back to the smoke filled rooms scenario of long ago.

I’ve been told that, given time, the system will level out in terms of it’s randomness of it’s adjustment of the ratings, how long would this normally take? Does it actually look at historical results and if so how long of a data base would you normally run with?

Regards,

Brian

7

Hi Jon,

That would be very helpful thank you, I’ll send through a file later today,

Regards,

Brian

8

Hi Brian,

I would be interested to see some examples of what you think is wrong. You mention a randomness of adjustment. There is no randomness, the adjustments are purely mathematical according to the RYA spec. so there will be an explanation. So we need to see what you think is wrong before I can advise you. The algorithm is such that all boats ratings will be gradually adjusted such that each boat has an equal chance of winning. If you have a particularly good sailor or their rating is biased then they may win the first few races but they won’t keep winning unless they are constantly improving. After a while the winners will be the boats that sailed better than they usually do.

With Sailwave if you use it in External/Debug mode with Excel you can see all the calculations.

If you want you can change the algorithms that are used.

Jon

···

On 22 August 2016 at 11:01, bmcelligott1@gmail.com [sailwave] sailwave@yahoogroups.com wrote:

Thanks for your input folks, it seems that in most of you seem to mange the system with very little manual intervention.

The only way of adjusting the median is by exporting the ratings and manually adjusting them and imputing them again? This to jimc_hjones point is back to the smoke filled rooms scenario of long ago.

I’ve been told that, given time, the system will level out in terms of it’s randomness of it’s adjustment of the ratings, how long would this normally take? Does it actually look at historical results and if so how long of a data base would you normally run with?

Regards,

Brian

Jon Eskdale
07530 112233

Skype “eskdale”