I understand (if I'm wrong, please correct me but refrain from
destroying my village) that, for a match to n points, the standard
time limit is to give each player 2 * n minutes for the match
(+ a per-move delay which I think is around ten seconds).
But isn't this too simplistic? Surely the relationship between
expected match length (in terms of total number of plays)
is non-linear in k where k is the minimum score needed to win
the match? Why don't the rules respect the complex (and perhaps
interesting) maths involved?
On 7/18/2023 1:13 PM, peps...@gmail.com wrote:
I understand (if I'm wrong, please correct me but refrain from
destroying my village) that, for a match to n points, the standard
time limit is to give each player 2 * n minutes for the match
(+ a per-move delay which I think is around ten seconds).
But isn't this too simplistic? Surely the relationship between
expected match length (in terms of total number of plays)
is non-linear in k where k is the minimum score needed to win
the match? Why don't the rules respect the complex (and perhaps
interesting) maths involved?
There's value in having simple rules that anyone can understand.
For a match to n points, the minimum number of games is 1 and the
maximum number is 2n-1. I don't know what the "average" length is,
but it is probably of O(n)
so the simple rule is probably fine.
On 7/18/2023 1:13 PM, peps...@gmail.com wrote:
expected match length (in terms of
total number of plays)
For a match to n points ..... I don't know
what the "average" length is, but .....
Surely the relationship between expected
match length (in terms of total number of
plays) is non-linear
in k where k is the minimum score needed
to win the match?
"ah....Clem" <ah_...@ymail.com> writes:
For a match to n points, the minimum number
of games is 1 and the maximum number is 2n-1.
I don't know what the "average" length is, but it
is probably of O(n)
Nice fundamental argument!
After looking at
https://www.bkgm.com/rgb/rgb.cgi?view+712
I just ran 20 matches each with GNU Backgammon
playing itself on expert level to augment the table.
Here is the average number of games for different
match lengths .....
is there any reason (theoretical or historical) for odd
match lengths?
Perhaps this is one of the reasons why 3 matches
to 7 might be more attractive than one 21-pointer.
I learned from Douglas Zare that the cube action for 7-point matches is
very similar to the cube action for money (more so than for 9-point
or 11-point matches, I believe), including the recubes (though certainly
not the re-recubes).
Which for me immediately raises the question which match lengths are particularly interesting in the sense of having many match scores at
which the cube action deviates strongly from money sessions. Also, is
there any reason (theoretical or historical) for odd match lengths?
As for odd match lengths, a surprising number of people will suggest...
that it's to avoid ties. This of course makes no logical sense, but
there are of course many contests in sports that have a "best-of-n"
form, where n needs to be odd to avoid ties. So perhaps people just
got used to match lengths being an odd number in other sports, and
carried over this tradition to backgammon even though there's no mathematical reason for it.
On 7/19/2023 11:51 AM, Axel Reichert wrote:
Which for me immediately raises the question which match lengths are particularly interesting in the sense of having many match scores atI learned from Douglas Zare that the cube action for 7-point matches is
which the cube action deviates strongly from money sessions. Also, is there any reason (theoretical or historical) for odd match lengths?
very similar to the cube action for money (more so than for 9-point
or 11-point matches, I believe), including the recubes (though certainly
not the re-recubes). I never confirmed this calculation myself, but
Zare is usually right about that sort of thing.
As for odd match lengths, a surprising number of people will suggest
that it's to avoid ties. This of course makes no logical sense, but
there are of course many contests in sports that have a "best-of-n"
form, where n needs to be odd to avoid ties. So perhaps people just
got used to match lengths being an odd number in other sports, and
carried over this tradition to backgammon even though there's no mathematical reason for it.
On Thursday, July 20, 2023 at 1:02:47 PM UTC+1, Timothy Chow wrote:
...
...
As for odd match lengths, a surprising number of people will suggest
that it's to avoid ties. This of course makes no logical sense, but
there are of course many contests in sports that have a "best-of-n"
form, where n needs to be odd to avoid ties. So perhaps people just
got used to match lengths being an odd number in other sports, and
carried over this tradition to backgammon even though there's no
mathematical reason for it.
I don't quite follow what you're saying here. For example, in mens pro tennis,
best of 3 sets and best of 5 sets are both common formats. And yes, both 3 and 5
are odd, and best of (for example) 6 sets wouldn't make any sense.
But what (some) people are puzzling over is why the winner of a backgammon match
is (almost) always first to 2 * n + 1 rather than first to 2n.
But first to 2n means best of 4n - 1 which is of course odd.
The thinking process you describe does nothing to explain why backgammon matches,
in contrast to other sports, are (almost) always best of 4n + 1 (for some n) rather than
best of 4n + 3. This can't be explained by pointing out that "best of k" needs k to be odd.
...
A former director of the lab where I work has a line which I love:
"Your problem, Tim, is that you're trying to use logic." I believe
that that applies here.
| Match length | Average Number of Games |
| 1 | 1.00 |
| 3 | 2.35 |
| 5 | 3.83 |
| 7 | 5.02 |
| 9 | 7.24 |
| 11 | 7.81 |
| 13 | 10.35 |
| 15 | 10.00 |
| 17 | 11.60 |
| 19 | 14.00 |
| 21 | 13.20 |
| 23 | 13.85 |
| 25 | 15.80 |
20 matches each is certainly to few, which
possibly explains the non-monotoneous growth.
But I can easily imagine, as Paul put it, some
nonlinear effects,
due to "overshooting" or "undershooting" the match
length with a particular (higher) cube level.
Which for me immediately raises the question which
match lengths are particularly interesting in the sense
of having many match scores at which the cube action
deviates strongly from money sessions.
. .... But I can easily imagine, as Paul put it,
some nonlinear effects, due to .....
On July 19, 2023 at 9:51:40 AM UTC-6, Axel Reichert wrote:
20 matches each is certainly to few, which
possibly explains the non-monotoneous growth.
What the heck "non-monotoneous" means? Can't you
bring yourself to say say non-linear? ;)
On 7/22/2023 5:37 AM, MK wrote:
On July 19, 2023 at 9:51:40 AM UTC-6, Axel Reichert wrote:
20 matches each is certainly to few, which
possibly explains the non-monotoneous growth.
What the heck "non-monotoneous" means? Can't you
bring yourself to say say non-linear? ;)
I think he meant "non-monotonic". https://en.wikipedia.org/wiki/Monotonic_function
So this terminology felt weird (and still does), the
backgammon way feels more natural.
As for odd match lengths, a surprising number of people will suggest
that it's to avoid ties. This of course makes no logical sense, but
there are of course many contests in sports that have a "best-of-n"
form, where n needs to be odd to avoid ties. So perhaps people just
got used to match lengths being an odd number in other sports, and
carried over this tradition to backgammon even though there's no
mathematical reason for it.
On 7/23/2023 4:35 AM, Axel Reichert wrote:
So this terminology felt weird (and still does), theNot only that, there's no reasonable way to describe a
backgammon way feels more natural.
backgammon match in "best-of-n" language, because one can
win or lose more than one point per game.
By the way, here's a puzzle that I think I have posted on rec.games.backgammon before, but which you may not have seen.
Two teams are playing in the World Series, which is a best-of-7
match ("4 Gewinnsaetze"). One team, the Slow Starters, always
loses the first game. The other team, the Late Chokers, always
loses *if* the series reaches a score of 3-3. Otherwise, the
two teams are evenly matched, and are equally likely to win any
particular game. Which team is more likely to win the series?
Timothy Chow <tchow...@yahoo.com> writes:
As for odd match lengths, a surprising number of people will suggestMight be. In Germany, the "best-of-n" wording came up only some decades
that it's to avoid ties. This of course makes no logical sense, but
there are of course many contests in sports that have a "best-of-n"
form, where n needs to be odd to avoid ties. So perhaps people just
got used to match lengths being an odd number in other sports, and
carried over this tradition to backgammon even though there's no mathematical reason for it.
ago, probably as an "Americanism". The standard wording for, say, a best-of-5 tennis match was "3 Gewinnsaetze", roughly translating as "the winner will need 3 won sets". I still remember being puzzled by the "best-of-n" way, because I needed to do calculations to come up with the needed number of won sets and also because in many cases not all n sets
are played. So this terminology felt weird (and still does), the
backgammon way feels more natural.
On 7/22/2023 5:37 AM, MK wrote:
On July 19, 2023 at 9:51:40 AM UTC-6, Axel Reichert wrote:
20 matches each is certainly to few, whichWhat the heck "non-monotoneous" means? Can't you
possibly explains the non-monotoneous growth.
bring yourself to say say non-linear? ;)
I think he meant "non-monotonic".
https://en.wikipedia.org/wiki/Monotonic_function
Timothy Chow <tchow...@yahoo.com> writes:
On 7/22/2023 5:37 AM, MK wrote:
On July 19, 2023 at 9:51:40 AM UTC-6, Axel Reichert wrote:
20 matches each is certainly to few, which
possibly explains the non-monotoneous growth.
What the heck "non-monotoneous" means? Can't
you bring yourself to say say non-linear? ;)
I think he meant "non-monotonic".
https://en.wikipedia.org/wiki/Monotonic_function
Yes, indeed. And of course, but at least you know
this, there is a difference between non-monotonic
and non-linear,
which is why I tried to use the former on purpose.
I have automated things with GNU Backgammon
(IMHO /the/ killer feature compared to XG)
and run 100 matches for all match lengths from 1
to 64 (the maximum allowed in GNU Backgammon).
Here are the results (a plot shows games and
moves to fit very nicely to straight lines),
which do not bear out the hypothesis that there
might be a different relation between the match
length and the number of games played than O(n):
So the number of games is pretty constant at
about 2/3 of the match length.
Even though longer matches would offer the the
opportunity for higher cubes, thus drastically
reducing the number of games played,
seems that the likelyhood of higher cubes (with
proper cube skill) diminishes faster than the match
length increases. See
https://www.bkgm.com/rgb/rgb.cgi?view+662
for how rare already a cube of 8 is.
I have done a similar thing for cubeless backgammon
(but only for matches up to 25 point),
The number of moves per game is again about 54,
no surprise here.
The number of games per match is pretty constant
again, this time at about 4/3 of the match length.
If we put this together we end up with
54 * 4/3 * m = 72 * m
42 * 2/3 * m = 28 * m
This of course does not amount to cubeless
backgammon being the more skillful game:
1. Cashing a boring (in the sense of low equity
difference between candidate plays) race or other
low-skill games cuts away the luck.
2. Imperfect human players might squander more
equity getting cube decisions wrong than checker plays.
As usual, in these things. "It doesn't matter which one you choose.
Both candies are exactly the same size!"
The late chokers start with a 1 0 lead.
They win the match if they win 3 of the next 5.
Because of the Tim-Termination condition, they lose if they win only
two of the next 5.
In a 50/50 context, winning at least 3 out of 5 is a 50/50 parlay.
So you can take either candy you want.
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 297 |
Nodes: | 16 (2 / 14) |
Uptime: | 17:34:32 |
Calls: | 6,667 |
Calls today: | 1 |
Files: | 12,216 |
Messages: | 5,336,860 |