On 12/07/2012 15:22, Don Dailey wrote:
I assumed with the log or square root function that the winner would get
this number of points and the loser would lose the same amount. And
it should be centered around a komi value in my opinion - so that if
komi is 5.5 you consider a score of 7.5 to be worth 2 points or sqrt(2).
I like your function just as well or even better. I'm not fussy about
the exact function used, just the arbitrary cap which has no logic at
all. Even if they had no cap it would be better.
You say it's not clear what they are trying to achieve, but I think that
was explained. The explanation left me reeling because it had to do
with strong programs crushing weak programs and not getting overly
rewarded for playing a weak opponent.
You can blame me for that, not Łukasz. He suggested the cutoff at
+-50, it seemed reasonable to me, and I tried to explain why.
But this system basically does
reward the program that plays the most weak opponents
This is irrelevant - the event is a round robin.
Nick
and I think it's
really flawed in general.
The real point of this is to impose a more western attitude to the game,
trying to "crush" your opponent - pick off every possible stone you
can, etc. In this system it's possible to lose most of your games
and still win the tournament because a small loss is hardly better than
a small win. With any of these alternative scoring systems that
situation is greatly improved.
I did think of one interesting format. Suppose that you use the
performance ratings of the programs in the tournament to decide the
winner, but you count each stone as a separate game? In 9x9 you have
81 games, so the score can be 81 - 0, or anything in between (I
strongly suggest centering this based on Komi.) A score of 81-0 is
impressive, but if the program is really weak it's performance rating
is going to be very low (not to mention the fact that beating it by this
score won't help its performance rating.) This has some of the
same flaws but it would still be fun. If you lose by a small margin
it's not much worse the winning by a small margin.
Don
On Thu, Jul 12, 2012 at 9:53 AM, Álvaro Begué <[email protected]
<mailto:[email protected]>> wrote:
We are dealing with a quantity (the score difference) which can be
positive or negative, so neither square root nor log are very natural
functions to use.
If you want diminishing returns, here's a better way to do it: Give
each player 1/(1+exp(-K*score)) points. As K goes to infinity this
converges to the usual loss=0, draw=1/2, win=1. As K goes to 0 it
becomes linear. Also the sum of points awarded to both players is 1,
which is kind of a nice property to maintain.
However, it is unclear what exactly they are trying to achieve, so I
can't really suggest a particular function to use.
On Thu, Jul 12, 2012 at 9:41 AM, Don Dailey <[email protected]
<mailto:[email protected]>> wrote:
> Like you I agree that they should not not use this at all.
However,
> given that this is what they want to do I would argue that using
using the
> square root function (or something like it) is really what they
want. If
> they believe 150 is no better than 50, but that it's suddenly
linear below
> 50, what sense does that make? It goes from being linear to being
> nothing. So using the log of the score, or the square root,
anything
> like that, is a huge win and gives them the basic behavior they
really want
> without having an arbitrary cap.
>
> And by the way, the square root is not just as arbitrary as
capping -
> unless you define arbitrary as anything you might do (which in a
sense it
> is.) So maybe arbitrary is not the right word here. What we
are looking
> for is something that is a more logical means to an ends - and
the cap is
> not nearly as good as the log or the square root.
>
> Don
>
>
>
> On Thu, Jul 12, 2012 at 9:25 AM, Álvaro Begué
<[email protected] <mailto:[email protected]>>
> wrote:
>>
>> The square root is just as arbitrary as capping at 50. The only
>> function I really like is capping at 0.5.
>>
>> Álvaro.
>>
>>
>> On Thu, Jul 12, 2012 at 9:16 AM, Don Dailey
<[email protected] <mailto:[email protected]>> wrote:
>> > Truncating to [-50 .. 50] seems rather arbitrary to me too.
There
>> > should
>> > either be no truncation at all, or if the concept is to not "over
>> > reward"
>> > big wins it should be replaced by a function such as the
square root of
>> > the
>> > score. In this way you get progressively less credit for
bigger and
>> > bigger wins.
>> >
>> > Don
>> >
>> >
>> > On Thu, Jul 12, 2012 at 5:48 AM, Rémi Coulom
<[email protected] <mailto:[email protected]>>
>> > wrote:
>> >>
>> >> Why truncate to [-50..50] ?
>> >>
>> >> On 10 juil. 2012, at 22:20, Łukasz Lew wrote:
>> >>
>> >> > Fellow Go enthusiasts,
>> >> > I would like you invite you to:
>> >> >
>> >> > Kaś Cup - a peculiar computer Go tournament.
>> >> >
>> >> > There will be prize pool of total 100$, yay!
>> >> > It will take place on 5th of August on KGS.
>> >> >
>> >> > The peculiarity will come from the scoring method.
>> >> > While this will be a Round Robin, the score for each game
won't be
>> >> > +-1
>> >> > point,
>> >> > but the exact result of the game truncated to the [-50 .. 50]
>> >> > interval.
>> >> >
>> >> > One last rule is that participants may not use more than 4
cores of
>> >> > CPU
>> >> > power.
>> >> >
>> >> > Nick kindly agreed to organize and look after the
tournament for
>> >> > which
>> >> > I am grateful.
>> >> > Also he is in charge of choosing a ruleset and time settings.
>> >> >
>> >> > Thank you and let us know if you will participate.
>> >> > Łukasz
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
--
Nick Wedd
[email protected]
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
