One point matches
Karl Juhnke
: WHAT IF it is intrinsically more difficult to play one point
: matches than to play games in which gammons count?
Interesting idea, but here is a counter hyposthesis: What if one-point
players on FIBS take advantage of our money instincts which don't apply
in one-point matches?
: Checker play decisions come up all the time in which you have
: to decide between different levels of aggressiveness. Very
: often a play which is too aggressive will lose more often
: but win more gammons in compensation.
Exactly. So when I see a possibility to gammon, I play aggressively.
This isn't warranted in one-point matches, but my positional judgement
has been built up in games where gammons count, so I have a hard time
adjusting.
Does this make sense?
Peace,
Fritz
Christian Glaser
In article <DEs1I...@world.std.com> w...@world.std.com (Walter G Trice) writes:
> I've been puzzled by the fact that individuals have achieved
> unusually high ratings on FIBS by limiting themselves to
> one point matches. Last night a possible explanation occurred
> to me.
>
that seems like a fact but maybe these players are just good.
another explanation is that they can play just more matches.
the same experience stands for many more matches.
> WHAT IF it is intrinsically more difficult to play one point
> like a bizarre idea, and I only want to argue that it MIGHT
> be true rather than that it actually IS true.
>
> Checker play decisions come up all the time in which you have
> to decide between different levels of aggressiveness. Very
> often a play which is too aggressive will lose more often
> a LOT of play problems where two very different plays result
> in about the same equity because more wins are neatly balanced
> by fewer gammons, or vice versa.
>
> But this never happens in one point match play. There
> will still be situations with balancing factors. A
> common example is the contact bearoff where the fact that a
> given play gets you hit more often is balanced by the fact that
> it results in your having extra checkers off when you get hit.
> Without gammons, though, perhaps different plays are less
> balanced on average and getting it right tends to be more
> important.
>
i think that their is no doubt about that the CUBE is a very important
part of the backgammon game as are gammons/backgammnons.
when you learn to play backgammon you grow up with this in mind
and adapt your strategy concerning gammons & cube decisions.
you realize that gammons are nice :) and go more aggressively for
it.
and then you suddenly face a situation with no cube & no gammon counting.
questions are arising:
do you know the now relevant new strategies, have you read the posting
from one_pointer, have you that experinence with this normally not
so frequent situation........
and that is the point were these individuals have an advantage:
even though the game without gammons & cube is less complex.
the normal player has less knowledge, experience, feeling & intuition
with this situation.
> Perhaps the stronger player rates to win more one point matches
> than three point matches against a given weaker opponent.
>
relative: i hope not, otherwise.....
absolute: yes, because they simply play more matches
(ok, if the weaker is a specialist for one-pointers who has
no idea of cube handling & match equties...)
> Can anyone refute this hypothesis?
>
> -- Walter Trice
just my 0.02dm :)
excuse the english
cheers
charles
--
__________________________________________
dlr oberpfaffenhofen, wt-dv-its
e-mail: christia...@dlr.de
tel: ++49+8153-28-2906
fax: -1136
David Montgomery
>I've been puzzled by the fact that individuals have achieved
>unusually high ratings on FIBS by limiting themselves to
I keep track of my FIBS results at different match
lengths versus players of different strengths, and I
have a little program that will iterate my results,
updating my rating, until it converges. My iterated
rating for one point matches has always converged to
over 2000.
I'm not sure why this is so. My iterated rating over
all matches has consistently averaged 1800-1825. This
difference is consistent with many hypotheses, namely:
0) I haven't gathered enough data.
1) My match strategy and cube handling skills are very weak.
(In light of your hypothesis I realize I should add:
my gammon handling skills are very weak.)
2) I play better relative to my opponents in the endgame
relative to the opening.
3) The rating formula does not work well for large rating
differences. (More of my one point matches are against
weaker players, for a variety of reasons.)
4) The rating formula underestimates the skill present in
one point matches.
Your hypothesis that one point matches might be intrinsically
more difficult than games with gammons technically would
not explain the high ratings of one point match players,
since if the rating formula accounted appropriately for
the skill in one point matches (whether or not they are
more difficult than games with gammons), then higher
ratings would not result. Presumably you are also
hypothesizing that 4) is true too.
My own belief has always been that 0) was a distinct
possibility, that 1) was probably true, but probably not to
the tune of 200-300 rating points, that 2) was relatively
insignificant, that 3) contributed but probably by less than
50 points (I base this on looking at my ratings iterated
separately against weaker players and stronger players),
and that 4) was the most likely candidate. With the advent
of one_pointer, loner and mloner I am more convinced than ever
that 4) is true.
With respect to the reasons that one point matches might
have more skill than you would expect, my intuition is that
the primary reason is that the games last longer. Although
cube decisions are more difficult than checker play decisions,
when you add an average of 25 extra checker plays, the
cumulative effect may be more important than one or two
primary cube decisions.
David Montgomery
monty on FIBS
Walter G Trice
unusually high ratings on FIBS by limiting themselves to
to me.
WHAT IF it is intrinsically more difficult to play one point
matches than to play games in which gammons count? This seems
like a bizarre idea, and I only want to argue that it MIGHT
be true rather than that it actually IS true.
Checker play decisions come up all the time in which you have
to decide between different levels of aggressiveness. Very
often a play which is too aggressive will lose more often
but win more gammons in compensation. In fact there must be
a LOT of play problems where two very different plays result
in about the same equity because more wins are neatly balanced
by fewer gammons, or vice versa.
But this never happens in one point match play. There
will still be situations with balancing factors. A
common example is the contact bearoff where the fact that a
given play gets you hit more often is balanced by the fact that
it results in your having extra checkers off when you get hit.
Without gammons, though, perhaps different plays are less
balanced on average and getting it right tends to be more
important.
Perhaps the stronger player rates to win more one point matches
than three point matches against a given weaker opponent.
Can anyone refute this hypothesis?
-- Walter Trice
Bernhard Kaiser
: -- Walter Trice
hi!
First i have to say, loner really cant be called an individual :)
i think there are three main reasons, why playing only onepointers could be
such fine to get a high rating.
1, Probably the rating formula does slightly underestimate the difference
between a good and a bad player compared to longer matches. (the same time
it slightly overestimates the difference in very long matches, i think).
2, Onepointers are mostly played until the end, without someone leaving.
Much points are lost when playing a longer match and your opponent quits,
just about to lose the match. Thats one reason more , why i like onies :)
I really havent much saved games.
3. (most important) Its indeed possible to specialize in onepointers.
of course there are much situation, where you have to chose different moves
than in money game or most match situations.
I really dont think, i would win more onepointers against a weaker opponent
than 3 pointers. Thats far too much.
But i try to play different against different rated players.
So i am , of course , just trying to find the objective best moves against
an expert, lik , say, kitwoolsey. Of course i really am not a 54 to 46
favorite against him, as the formula indicates. Probably i am not even
the favorite.
But i think, my advantage is, i have a pretty good feeling how to play weak or
very weak opponents .
against a very weak opp. i often try out moves, which differs considerably from
the "best" move. And, believe me, i am coming through with it.
(for example, against a weak opponent ,playing 13 7 13 11 with an opening 6-2
is much better than the objective correct 24 18 13 11, in my feelings)
Someone could ask, why loner still is better than me, although he always plays
the same, against every opponent.
Well, its because his play really is outstanding! he makes , if at all ,
only very slight errors; he doesnt get tired, and doesnt run hot :)
onepointer
Farhan Malik
>Someone could ask, why loner still is better than me, although he always plays
>the same, against every opponent.
>Well, its because his play really is outstanding! he makes , if at all ,
>only very slight errors; he doesnt get tired, and doesnt run hot :)
>onepointer
But is loner really better than one-pointer? one-pointer has
a rating of 2000 with just over 1000 games experience. Loner has over
5000 games experience but it's rating is only marginally higher than
one-pointer's. It seems that one-pointer's winning percentage is much
better than loner's and that it is only a matter of time before
one-pointer passes it up.
The only real way to settle this is a duel. 20 one point
matches to take place on the observation deck of the world trade
center. Loner on one side, one-pointer on the other. Buy your
tickets now.
Farhan
beers
I really dont know, but I have a question;
Aren't one-pointers just easier to play due to the fact they are less
complex than x-point games ? Thus having less chance of making errors ?
In, for example three-in-a-row its easy to become an expert and never lose.
I think I lost quite some matches due to lack of good doubling skills..,
Maybe I would have won those in plain no-cubing matches.
?
Since I cant find this argument I'm probably wrong, but why ?
beers
Ron Karr
It's true that doubling adds a whole extra dimension to the game, and so from an absolute point of view, I believe that multiple-poi=
nt matches do require more skill than one-point matches.
Certainly a player with good doubling skills will have a big edge over one without them. However, in most cases, players with super=
ior doubling skills will also have superior checker skills. Most backgammon players with a fair amount of experience are quite accu=
stomed to playing with the doubling cube. They're used to playing in situations where the cube is in effect, and where gammons coun=
t (money games, and most match situations). Their strategy has been learned playing these types of games.
If you then match one of these players (say myself) against a player (or computer) who is an expert at playing one-point matches, su=
ddenly I find myself at a disadvantage. Yes, I'm aware that gammons don't count and there's no cube, so I will adjust in certain si=
tuations. Still there are a great many situations where I will instinctively make a play which is clearly correct when playing for =
money, but which is inferior in a 1-point match. For example, suppose there's a play which results in 30% gammons and 30% plain win=
s, and another which results in no gammons and 62% plain wins. The first play is correct by a huge margin when playing for money, w=
hereas the second would be correct in a 1-point match. Most players would be hard pressed to find that second play; loner would fin=
d it easily.
Ron
Karl Juhnke
: i think there are three main reasons, why playing only onepointers could be
: such fine to get a high rating.
Thank you very much for your first-hand opinions!
: 1, Probably the rating formula does slightly underestimate the difference
: between a good and a bad player compared to longer matches. (the same time
: it slightly overestimates the difference in very long matches, i think).
Exactly, overestimating the one is the same as underestimating the other.
: 2, Onepointers are mostly played until the end, without someone leaving.
: Much points are lost when playing a longer match and your opponent quits,
: just about to lose the match. Thats one reason more , why i like onies :)
: I really havent much saved games.
: 3. (most important) Its indeed possible to specialize in onepointers.
: of course there are much situation, where you have to chose different moves
<snip>
: But i think, my advantage is, i have a pretty good feeling how to play weak or
: very weak opponents .
: against a very weak opp. i often try out moves, which differs considerably from
: the "best" move. And, believe me, i am coming through with it.
: (for example, against a weak opponent ,playing 13 7 13 11 with an opening 6-2
: is much better than the objective correct 24 18 13 11, in my feelings)
This could be put as a reason unto itself, i.e. #4, or probably #1 as it
may be the most important. :) For those of you who frequent the chess
server ICC as well, I wonder whether Roman hasn't pushed his rating up
partly by knowing how to hustle "mere" masters and experts, as opposed to
being able to outduel other grandmasters.
Thanks again, onepointer, for your explanations.
Peace,
Fritz
Walter Swan Jr
In my opinion it is much easier for an expert to achieve a higher rating
in one point matches. This is becuase all games are played to a
conclusion, therefore the expert has the entire series of rolls to
outplay an lesser player. WIth the cube, the sequence of rolls is cut short.
It is much easier for a weaker player to snatch defeat out of the claws
of victory in a one pointer.
Walter G Trice (w...@world.std.com) wrote:
: I've been puzzled by the fact that individuals have achieved
: unusually high ratings on FIBS by limiting themselves to
: one point matches. Last night a possible explanation occurred
: to me.
: WHAT IF it is intrinsically more difficult to play one point
: matches than to play games in which gammons count? This seems
: like a bizarre idea, and I only want to argue that it MIGHT
: be true rather than that it actually IS true.
--
_
( '
walt\___)w...@netcom.com
~~~~~~~~~~~~~~~~~~~~~~~~~
Kit Woolsey
work out better for the stronger player (rating-wise) than longer matches
is the computer loner. When playing one-point matches, loner appears to
have stablized with a rating of around 2000. However, mloner's true
rating appears to be somewhere between 1900 and 1925. This is exactly
the same playing program, except that mloner uses the cube. As anyone
who has played mloner several matches can affirm, mloner's cube handling
in matches is nothing less than superb, so there is no major weakness
there. Therefore, the inevitable conclusion about why loner's rating is
so much higher than mloner's must be that loner plays only one-point
matches, so, whatever the reason, it is clear that one-point matches are
biased more favorably to the stronger player than longer matches
according to the current FIBS rating formula.
Kit
Barry J. Miller
point match rates to go on for many more moves than a normal game with
the cube available. Accordingly, the weaker opponent has that many more
opportunities to make errors which the stronger player will exploit -
errors which may never occur in a "normal" game because one side or the
other would have dropped a double or redouble.
Barry Miller
bjm on FIBS
Bernhard Kaiser
: The only real way to settle this is a duel. 20 one point
: matches to take place on the observation deck of the world trade
: center. Loner on one side, one-pointer on the other. Buy your
: tickets now.
: Farhan
Nice idea! who will pay me my tickets ?
BTW. i already are playing a series of best of 19 against loner,
the winning price is 100 $. Acyually i am leadung 4-3
onepointer
Bernhard Kaiser
: One very strong piece of evidence we have that one-point matches tend to
: Kit
Of course you are right, Kit!
But i still dont think, that the cube actions of mloner are as perfect
as the movements are, since it must be much more difficult to evaluate
for a computer.
I watched mloner a long time; and although he really doubled nice, there
were DEFINITELY a few cube errors, what i cant say from any moves.
Also dont forget my second point! Longer matches are much more often
leaved by the trailer. That could also make some difference in the rating.
Anyway, its probably not a big deal, and of course playing onepointers only
must be favorable, since i am for sure not better than anyone in the top 20.
onepointer