I would second the general direction of Jim C's thoughts although I
wouldn't go quite as far he does. I think that in some common situations
there can be a significant difference between scoring systems.
If essentially the same boats sail in all of the races of a series then
for the most part I think you'll find that all scoring systems come up
with pretty much the same result (and most of them will come up with
exactly the same result). The differences occur when you have a series
of races longer than a regatta (in other words, a series not on
consecutive days like a once-per-week season championship) and the
number of boats varies from race to race. Even there the differences
won't be gigantic although the policy signals they send might be important.
A system that uses some kind of averaging can have the advantage of
being "fairer" but that comes with the disadvantage of being harder for
competitors to understand and creating incentives for some boats to stop
sailing in the last few races (since poor results in the last few races
will bring down your average and worsen your score). A system that
simply sums your scores and discards your worst has the disadvantage of
being slightly less "fair" (at least in theory) but has the advantage of
being easy to understand and creating an incentive to keep coming out in
the later races of a series since you can only improve your placement in
the series. You'll have to decide for yourself which goal is more
important - personally, I lean toward sacrificing a tiny bit of
theoretical fairness in favor of competitor understanding and incentives
to keep coming out racing.
I tend to think the "Percentage Point System" is probably the fairest of
all. The points for each race are a simple high point (boats beat + 1).
Your score at the end is a percentage with your points divided by the
points if you'd won ever race you entered. 100% means you won every race
you entered while 50% means that on average you finished in the middle
of the fleet in every race. You have to set some minimum number of races
a boat must enter to be eligible for series trophies at the end. Now the
disadvantages of such a system should be fairly obvious - it is kind of
hard for competitors to wrap their heads around where they stand and
what they need to do in future races. Plus, once you've entered the
minimum number of races you risk having your "average" go lower with a
bad result, creating some potential incentive not to race in the last
few races.
My club thought the disadvantages of the Percentage Point System
outweighed the advantages BUT we really liked the "percentage" concept.
So, we took the percentage concept and apply it to each race so the per
race score is essentially a percentage. We then sum and discard races in
a traditional way. Sailors tend to be a bit mystified as to how the
points for a race are generated but at least the incentive to keep
coming out in future races is clear and unambiguous. You'll see some
description of that system in the User Group as the "Sunset System."
As Jim C points out, it probably doesn't make much difference which of
many high-point type systems you use. They all have their own
characteristics and reward slightly different things - the only really
important think is to recognize that the default provisions in the
rulebook (rule A9) are widely considered to be totally unfair and should
be avoided at all costs unless you have the highly unusual situation
where every boat sails in, and finishes, every race of the series.
WHAT TO LOOK FOR
Fixed vs Variable Gaps. 50+ years ago it was popular to have the gap
between places vary on the theory that in a large fleet it was easier to
move from 50th to 49th then from 2nd to 1st. In the older rulebooks that
was called either the "Bonus Point" or "Olympic" system (it was used for
years to score Olympic sailing). That approach has been mostly abandoned
in recent decades although a few of the high-point scoring systems still
incorporate that approach. You'll have to decide whether you prefer a
"fixed gap" or "variable gap" system. Personally, I tend to favor the
modern trend toward "fixed gap" systems.
Large vs Small Fleets. Some scoring systems assume that the value of
races should vary based on the number of competitors with some systems
favoring races with more competitors while other systems favoring races
with fewer competitors. The logic for favoring races with more
competitors is that it is harder to win in a large fleet while the logic
for favoring races with fewer competitors is that when the weather gets
ugly only the best competitors come out so in a smaller fleet you are
competing against only the best competitors so those races should count
more.
Personally, I find both arguments somewhat persuasive and think that
each is probably going to be true about half the time. So, in my view
you are better off with a system that doesn't have a bias based on the
number of boats in a race since such an unbiased system will be equally
accurate all of the time. CHIPS has a bias toward boats that sail in
smaller classes. Rinderle B tends to favor boats that sail in larger
classes but has some distortions in the bottom half of the fleet.
Cox-Sprague favors boats in larger classes. The Sunset system is
indifferent to class size (a middle boat gets the same points regardless
of class size).
Incentives to Keep Coming Out. Systems that use averaging can create
incentives to not come out for races at the end of a series. All
averaging systems have these incentives to varying degrees - including
the Batting Average system referred to in an earlier posting,
Cox-Sprague and the Percentage Point system referred to above. If you
are going to use such a system I'd recommend setting the "minimum"
number of races relatively high so that boats need to keep coming out to
qualify for the trophies. [Under a pure averaging system with no minimum
the boat that wins the first race has no incentive to ever come out a
second time since their average can only go down!]
Art
PS - I have a half-written paper on my computer about the issues that
arise in picking a scoring system for a long series of races. Perhaps my
New Years resolution should be to finish that and post it in the
documents section of the user group!
···
On 11/24/2016 2:15 AM, yho@devboats.co.uk [sailwave] wrote:
---In sailwave@yahoogroups.com, <tdh_sails_65@...> wrote :
Now that I am past the visual concerns of reporting scores, I am
actually looking at the scores themselves and thinking about if
things are working as we would like.
before you go very far down that road, I suggest you take the results
from a typical past series of your club, get them into a spreadsheet
and model what changes you'll get. I got very excited about the
concept of better scoring systems a few years ago, dreamed up what I
thought would be a vastly improved method (combining features of high
point and low point scoring) wrote a spreadsheet to automatically
calculate it and...
...discovered it made very little difference to our club results at
all. I then tried a whole number of other systems, and discovered
that while the odd place or two might swap round, fundamentally with
our results it didn't much matter what scoring system I used, So I
abandoned my custom spreadsheet in disgust. The only exception was a
very odd system called ostrobogolous or ostro, which did what was
intended in keeping a series alive to the end, but on close
examination I found that it did this by weighting the series towards
the later races to such an extent that it didn't matter if a
competitor came first or last in the first race, their final points
total was exactly the same.
Now your club results and experience may very well be different from
mine, it probably depends on attendance pattern. So all I'm
suggesting is that before you go very far down the road of how you
will calculate results with a different scoring system you first
model what differences it will make, and decide whether its worth the
effort.
Jim C
