"David Richerby" <davidr@[EMAIL PROTECTED]
> wrote in message
news:oVD*HPw+r@[EMAIL PROTECTED]
> David Kane <davidekane@[EMAIL PROTECTED]
> wrote:
>> "David Richerby" <davidr@[EMAIL PROTECTED]
> wrote:
>>> David Kane <davidekane@[EMAIL PROTECTED]
> wrote:
>>>> In tournaments that have used alternate scoring, there have been
>>>> no accusations of cheating. There were, however, a higher than
>>>> normal percentage of highly contested games. Inconvenient facts,
>>>> I know.
>>>
>>> Please cite a high-level chess tournament that has used scoring
>>> other than a draw being worth half as much as a win, to each
>>> player. (I am not aware of any having been played.) For bonus
>>> marks, please cite enough of them that we have a statistically
>>> significant number of games to consider.
>>
>> The GM Slugfest is the most famous that I'm aware of. It used BAP
>> scoring. (Black win =3, White win =2, Black draw =1)
>
> OK, that's a start. From what I can see, there were two of these
> tournaments, one in 2005 and one in 2006. According to [1], the 2006
> tournament was held over three days and had a field of fourteen
> players whose strength ranged from 2670 down to 1952 (USCF), with the
> following spread:
>
> 2600+ 5 players
> 2500-2599 2
> 2400-2499 2
> 2300-2399 1
> 2200-2299 3
> 2100-2199
> 2000-2099
> 1900-1999 1
>
> I've not found a full crosstable anywhere so I don't actually know
> what percentage of the games in the tournaments were draws. However,
> the very wide spread of abilities among the players would tend to
> reduce the frequency of draws under any scoring system, as would the
> pressures of a short Swiss tournament, where the only way to win the
> tournament is to win almost all your games. The wide range of
> abilities in that tournament seems very significant to me.
>
>
> I'd also say that the BAP system is extremely unfair. In particular,
> in the given tournament, we can probably assume that the 1952-rated
> player lost all his games, since he was the weakest player by nearly
> 300 points[2]. Half of the people who played against him were given
> the white pieces so their `free wins' were worth only two points; the
> other players were given three points for their wins against him. Mig
> Greengard[3] and GM Joel Benjamin[4] don't think BAP is a good idea,
> especially in tournaments with a wide spread of player strength.
A few points:
The normal scoring system is biased toward White. BAP is biased toward
Black (though the bias is less than White's with normal scoring) There is
nothing fundamentally different about the two cases. Technically you are
correct, however, that it is "unfair". Similarly normal scoring is
"unfair" yet
it hasn't prevented it from being used.
In any Swiss tournament, the bottom player will almost never play the top
players.
The 1900 player I believe was Clyde Ballard (the "B" in BAP), included to
give
the
experimental tournament an even number of players.
The "beliefs" held by random folks aren't im****tant; the strengths of
their
arguments are.
>
> [1] http://chesslodge.blogspot.com/2006/10/gm-slugfest-tournament.html
> [2] That is, every other player in the field would expect to win at
> least 85% of their games against him and the 2600+ players would
> be expected to win over 99% of the time.
> [3] http://www.chessninja.com/dailydirt/2005/11/slugfest_chess.htm
> [4] http://main.uschess.org/content/view/6741/341
>
>> BTW, how does dreaming up ludicrous cheating scenarios help create a
>> statistically significant number of games?
>
> It is, as I have pointed out before, unreasonably of you to suggest
> that I am in some way deficient for failing to provide statistical
> evidence of the effect of something that has never been tried.
>
I never requested statistical evidence. But the fact that you
haven't provided *any* evidence is telling.
>> You sound like someone who is desperate for these experiments not to
>> go forth, lest they prove you wrong. Seriously, why is that? Why
>> are you so attached to the status quo that you are willing to make
>> irrational and false arguments to preserve it?
>
> Please stop trying to side-track and personalize the discussion by
> casting aspersions as to my motives. Stop for a moment and see if
> your arguments are in any way specific to the matter at hand and you
> will see that they are not. That is a sure sign that something is
> going wrong.
I think motives are im****tant. I suspect that you really
are not so disconnected from reality as to believe your
argument. More careful consideration on your part, including
an examination of your motives, might let you see the merits
of my arguments as well as put forth sounder objections to them.
>
> You have proposed a change to the laws of chess. I have suggested
> that this change has flaws. You accuse me of irrationality and fear
> of change rather than substantively rebutting what I am saying.
I have not proposed changes to the laws of chess, but to scoring
which is external to the game itself. You have not simply
suggested that the proposed changes have flaws (which
would be correct), but claimed a specific flaw, which I believe
I have characterized fairly, as well as have rebutted.
>
>> I'm the first to admit that whatever experiments are tried are
>> bound to have glitches and might not bring the intended
>> consequences.
>
> On the one hand, you claim to accept that your idea might not work.
> On the other hand, anybody who gives an actual reason why your idea
> might not work (it makes cheating more profitable than the current
> rules) is presenting `irrational and false arguments'.
>
>> I would rather see [...] chess played under a mathematically sounder
>> alternate scoring system.
>
> Please justify your assertion that awarding less than half a point for
> a draw is mathematically more sound.
I suppose it is not pure mathematics, but empirically the half point
draw has produced a situation where contestants routinely
avoid playing real games. Imagine, for example, we awarded a
draw the value of 0.9. What do you predict that change would
make to the game, and would it improve the game? Same
arguments apply to 0.5.
|
