You should really provide an XG ID or GNU ID when posting here. Certain people of which I know one well won't even look at the problem otherwise. (because of how it's displayed)
Stick
Note the score, it's Gammon-Go for X. Clockwise play (of course!).
Score is X:5 O:6 7 pt.(s) match.
+24-23-22-21-20-19------18-17-16-15-14-13-+
| X X X O O O | | O X |
| X X X O O O | | O X |
| X | | |
| | | |
| | X | |
| |BAR| |
| | | |
| | | O |
| | | O |
| X | | X O |
| O O X | | O X X O |
+-1--2--3--4--5--6-------7--8--9-10-11-12-+
Pip count X: 250 O: 158 X-O: 5-6/7 Crawford
X to play 53
--
Zorba
On Monday, January 3, 2022 at 1:52:41 PM UTC-5, Zorba wrote:
Note the score, it's Gammon-Go for X. Clockwise play (of course!).
Score is X:5 O:6 7 pt.(s) match.
+24-23-22-21-20-19------18-17-16-15-14-13-+
| X X X O O O | | O X |
| X X X O O O | | O X |
| X | | |
| | | |
| | X | |
| |BAR| |
| | | |
| | | O |
| | | O |
| X | | X O |
| O O X | | O X X O |
+-1--2--3--4--5--6-------7--8--9-10-11-12-+
Pip count X: 250 O: 158 X-O: 5-6/7 Crawford
X to play 53
--
Zorba
You should really provide an XG ID or GNU ID when posting here.
Certain people of which I know one well won't even look at the problem otherwise. (because of how it's displayed)
Note the score, it's Gammon-Go for X. Clockwise play (of course!).
Score is X:5 O:6 7 pt.(s) match.
+24-23-22-21-20-19------18-17-16-15-14-13-+
| X X X O O O | | O X |
| X X X O O O | | O X |
| X | | |
| | | |
| | X | |
| |BAR| |
| | | |
| | | O |
| | | O |
| X | | X O |
| O O X | | O X X O |
+-1--2--3--4--5--6-------7--8--9-10-11-12-+
Pip count X: 250 O: 158 X-O: 5-6/7 Crawford
X to play 53
I'll play 23/18 but with no analysis, confidence, conviction, or poetry,
like a blind man in a dark room looking for a black cat that isn't there.
Stick Rice writes:
You should really provide an XG ID or GNU ID when posting here.
Certain people of which I know one well won't even look at the problem
otherwise. (because of how it's displayed)
... which happens only after your reply. Zorba's original post still was perferctly fine. d-:
On 1/3/2022 3:36 PM, Axel Reichert wrote:
Stick Rice writes:
You should really provide an XG ID or GNU ID when posting here.
Certain people of which I know one well won't even look at the problem
otherwise. (because of how it's displayed)
... which happens only after your reply. Zorba's original post still was perferctly fine. d-:My understanding is that Stick doesn't like ASCII art even when
it's properly displayed.
---
Tim Chow
It's a close call between the first two plays and the hit is not that
far behind.
XG's reliability for such a position (8 backcheckers) is probably not up
to its usual quality.
1. Rollout¹ Bar/22 9/4 eq:-0,174
Player: 38,03% (G:3,29% B:0,37%)
Opponent: 61,97% (G:44,28% B:21,04%)
Confidence: ±0,008 (-0,182..-0,165) - [84,2%]
Duration: 8 hours 38 minutes
2. Rollout¹ Bar/22 23/18 eq:-0,180 (-0,006)
Player: 37,87% (G:3,14% B:0,27%)
Opponent: 62,13% (G:45,82% B:23,66%)
Confidence: ±0,009 (-0,189..-0,171) - [15,8%]
Duration: 7 hours 49 minutes
3. Rollout¹ Bar/22 8/3* eq:-0,205 (-0,031)
Player: 37,10% (G:2,68% B:0,22%)
Opponent: 62,90% (G:47,27% B:25,56%)
Confidence: ±0,011 (-0,216..-0,193) - [0,0%]
Duration: 7 hours 57 minutes
¹ 11664 Games rolled with Variance Reduction.
Dice Seed: 48563399
Cubeless
First 6 moves: 4-ply
Remaining moves: 3-ply
Search interval: Large
XGID=---a-aBaBA--dB---b-bbbBBCA:0:0:1:53:5:6:1:7:10
Score is X:5 O:6 7 pt.(s) match.
+24-23-22-21-20-19------18-17-16-15-14-13-+
| X X X O O O | | O X |
| X X X O O O | | O X |
| X | | |
| | | |
| | X | |
| |BAR| |
| | | |
| | | O |
| | | O |
| X | | X O |
| O O X | | O X X O |
+-1--2--3--4--5--6-------7--8--9-10-11-12-+
Pip count X: 250 O: 158 X-O: 5-6/7 Crawford
Cube: 1
X to play 53
The title doesn't fit the post.
Here, you can't lose much equity, no matter what you do
so the back game player doesn't really need to worry.
It's a close call between the first two plays and the hit is not that
far behind.
XG's reliability for such a position (8 backcheckers) is probably not up
to its usual quality.
1. Rollout¹ Bar/22 9/4 eq:-0,174
Player: 38,03% (G:3,29% B:0,37%)
Opponent: 61,97% (G:44,28% B:21,04%)
Confidence: ±0,008 (-0,182..-0,165) - [84,2%]
Duration: 8 hours 38 minutes
2. Rollout¹ Bar/22 23/18 eq:-0,180 (-0,006)
Player: 37,87% (G:3,14% B:0,27%)
Opponent: 62,13% (G:45,82% B:23,66%)
Confidence: ±0,009 (-0,189..-0,171) - [15,8%]
Duration: 7 hours 49 minutes
3. Rollout¹ Bar/22 8/3* eq:-0,205 (-0,031)
Player: 37,10% (G:2,68% B:0,22%)
Opponent: 62,90% (G:47,27% B:25,56%)
Confidence: ±0,011 (-0,216..-0,193) - [0,0%]
Duration: 7 hours 57 minutes
¹ 11664 Games rolled with Variance Reduction.
Dice Seed: 48563399
Cubeless
First 6 moves: 4-ply
Remaining moves: 3-ply
Search interval: Large
XGID=---a-aBaBA--dB---b-bbbBBCA:0:0:1:53:5:6:1:7:10
Score is X:5 O:6 7 pt.(s) match.
+24-23-22-21-20-19------18-17-16-15-14-13-+
| X X X O O O | | O X |
| X X X O O O | | O X |
| X | | |
| | | |
| | X | |
| |BAR| |
| | | |
| | | O |
| | | O |
| X | | X O |
| O O X | | O X X O |
+-1--2--3--4--5--6-------7--8--9-10-11-12-+
Pip count X: 250 O: 158 X-O: 5-6/7 Crawford
Cube: 1
X to play 53
eXtreme Gammon Version: 2.10, MET: Kazaross XG2
--
Zorba
Glad you put on a hefty rollout. My rollout had the other two plays as whoppers!
Stick
On 1/9/2022 12:13 PM, J R wrote:
Glad you put on a hefty rollout. My rollout had the other two plays as whoppers!
StickYou're giving Murat a run for his money in terms of humor value!
---
Tim Chow
Why don't you rewrite the definition of whopper for everyone so that it makes sense then and submit it to bkgm.com.
On 1/10/2022 2:02 AM, J R wrote:
Why don't you rewrite the definition of whopper for everyone so that it makes sense then and submit it to bkgm.com.The definition as it stands is fine. What's at issue are the
conditions under which one is justified in *calling* something
a whopper.
I'm inferring that you're operating under something like the
following conditions: One is justified in calling a play a whopper
if and only if the results of a rollout (by a reputable backgammon
bot, with rollout settings that are generally considered to be
adequate) have been exhibited, and the EMG deficit of the play in
question (as compared to the top play of the rollout) as estimated
and displayed by the bot, is at least 0.1.
Notice that under these conditions, it's still possible to make
a "mistake" in the sense of *calling* something a whopper when it
isn't. The bkgm.com definition of a whopper refers to the *actual*
EMG deficit, under the assumption that equities exist and that
both players adopt an equilibrium strategy (commonly known as
"perfect play"). The actual EMG deficit in the vast majority of
situations is unknown, because backgammon has not been solved. So
whether something is actually a whopper is also unknown. Moreover,
even if we accept a hypothetical "infinitely long rollout" as a
surrogate for actual perfect play, we still don't have direct access
to it; all we have are finite rollouts. When one calls something
a whopper under the above conditions, one is making a judgment call
that the evidence is "good enough" to justify the claim.
The only difference between my operating conditions and yours is
that at the end of the last sentence, I say, "is at least 0.1, or
is very close to 0.1." The reason I do this is, in part, to remind
people of the above approximations and uncertainties. We use words
like "whopper" to convey approximate judgments; while we define them
in terms of numerical criteria for the sake of concreteness, it's
always a good idea to remind ourselves not to fall into the trap of
false precision.
For comparison, let's look at Ballard and Weaver's book on the
openings. They define terms like "very close" and "marginally
correct" and "blowout" and so forth, and then in the text, they
use only these words and don't display rollout results. Behind
the scenes, they're doing some rollouts and using certain cutoff
values, but they're deliberately hiding the rollouts because they know
that all we can realistically do is talk in terms of approximate
judgments, and they don't want readers to get hung up on whether
some rollout did or did not cross some arbitrary numerical borderline.
---
Tim Chow
Your argument is a whopper.
On 1/11/2022 12:24 PM, J R wrote:
Your argument is a whopper.Sorry you can't understand it. I tried to make it as
simple as possible.
Glad you put on a hefty rollout. My rollout had the other two plays as whoppers!I think those are called Whoppers with approximate cheese.
Stick
Some alternatives are:
1) He feels a psychological need to insult you.
2) He has spotted an error in your reasoning, that has eluded both you and me, but doesn't want to invest the
time in explaining it.
3) You are/were making some error or misconception in some related area of backgammon, that isn't 100%
relevant to your point here. He (rightly) objects to this error but won't point it out because he knows it's not fully relevant.
I'd vote for 3) because I've seen this type of Stick and Tim discussion before.
Tim: Adages like "make the 5 point...", "make the DMP play..." etc. are not helpful (except to beginners of course).
Stick: Well, you're obviously doing something wrong because you're not a good player.
On 1/12/2022 6:32 AM, peps...@gmail.com wrote:
Some alternatives are:
1) He feels a psychological need to insult you.
2) He has spotted an error in your reasoning, that has eluded both you and me, but doesn't want to invest the
time in explaining it.
3) You are/were making some error or misconception in some related area of backgammon, that isn't 100%
relevant to your point here. He (rightly) objects to this error but won't point it out because he knows it's not fully relevant.
I'd vote for 3) because I've seen this type of Stick and Tim discussion before.The trouble with 3) is that the present discussion has nothing to do
Tim: Adages like "make the 5 point...", "make the DMP play..." etc. are not helpful (except to beginners of course).
Stick: Well, you're obviously doing something wrong because you're not a good player.
with backgammon strategy. It's purely about linguistics and numbers.
I do think that the distinction between what something is in actuality (metaphysics) and our justification for calling it that (epistemology)
is highly relevant to the discussion, but it's a subtle distinction
that is difficult to grasp. You've mentioned Gettier counterexamples
before, ...
Stick's behavior is akin to that of a policeman who tickets a driver
for going 57 mph when the speed limit is 55 mph.
On 13-1-2022 15:54, Timothy Chow wrote:
Stick's behavior is akin to that of a policeman who tickets a driver
for going 57 mph when the speed limit is 55 mph.
As compared to traffic cop T. C. who epistemologically tickets drivers
for going 53mph when the speed limit is 55mph.
On 1/13/2022 1:48 PM, Zorba wrote:
On 13-1-2022 15:54, Timothy Chow wrote:
Stick's behavior is akin to that of a policeman who tickets a driver
for going 57 mph when the speed limit is 55 mph.
As compared to traffic cop T. C. who epistemologically tickets drivers
for going 53mph when the speed limit is 55mph.
Say what? You're saying that I have complained when people call
something with a displayed rollout value of > .1 a whopper?
---
Tim Chow
On 14-1-2022 14:51, Timothy Chow wrote:
On 1/13/2022 1:48 PM, Zorba wrote:
On 13-1-2022 15:54, Timothy Chow wrote:
Stick's behavior is akin to that of a policeman who tickets a driver
for going 57 mph when the speed limit is 55 mph.
As compared to traffic cop T. C. who epistemologically tickets
drivers for going 53mph when the speed limit is 55mph.
Say what? You're saying that I have complained when people call
something with a displayed rollout value of > .1 a whopper?
---
Tim Chow
No, I'm saying you call something with a calculated rollout value that's
less than 0.1 worse than the best move in the rollout, a whopper.
That's like the traffic cop giving you a ticket when you're not
speeding, but "almost" speeding and then coming up with a kind of story
about how his laser gun might be inaccurate.
Suppose a coach is trying to
correct a tennis player's tendency to serve too wide. If, during
training, the coach occasionally says that a serve is two feet past
the line when it's only 23 inches past the line, then should the
player fire the coach for not knowing the definition of a foot? Is
the coach's behavior comparable to that of the police officer in your analogy?
On 16-1-2022 19:26, Timothy Chow wrote:
Suppose a coach is trying to
correct a tennis player's tendency to serve too wide. If, during
training, the coach occasionally says that a serve is two feet past
the line when it's only 23 inches past the line, then should the
player fire the coach for not knowing the definition of a foot? Is
the coach's behavior comparable to that of the police officer in your
analogy?
If that coach used an electronic device that showed the distance as off
by less than two feet, it would be quite strange for him to insist that
it's actually two feet.
The coach examines an electronic device which says 23.5 inches, plus or
minus an inch. He yells to the player, "That was way out! Like two
feet!"
The player walks over and examines the electronic device. "That doesn't
say two feet; it says 23.5 inches. Two feet is 24 inches."
"When I said two feet, I
meant approximately two feet."
And actually, if you want to quibble
about it, the device itself says 23.5 plus or minus 1 inch, so 24
inches is even within the error bars.
Anyway, the point is that your
serve is way wide."
The player replies, "That's the saddest thing I've ever heard you say.
Maybe you should write to the National Institute of Standards and
Technology and tell them that their definition of a foot is wrong."
"Don't you know what it means to use approximate language?"
Etc. Do you think the coach's behavior is unreasonable here?
A whopper is defined as a range upwards from a threshold value, unlike a foot.
On 1/17/2022 4:01 PM, Zorba wrote:
A whopper is defined as a range upwards from a threshold value, unlike
a foot.
If this were true, I could see your point. Checking bkgm.com again, I
see that it agrees with you, but it doesn't match how I've seen the term used.
On 17-1-2022 15:02, Timothy Chow wrote:
The coach examines an electronic device which says 23.5 inches, plus or minus an inch. He yells to the player, "That was way out! Like twoI like that you take the issue as serious as it deserves to be taken,
feet!"
haha! Anyway.
In backgammon, calling something a whopper means it is *more* than a 0.1 error.
So the correct analogy would be the coach saying "that was more than two
feet out!".
The player walks over and examines the electronic device. "That doesn't say two feet; it says 23.5 inches. Two feet is 24 inches."So the device shows *less* than two feet, not *more*. The player is
right to complain.
"When I said two feet, IHe meant less than two feet in this case.
meant approximately two feet."
And actually, if you want to quibbleSo it *could* have been more than two feet out, and if we suppose some
about it, the device itself says 23.5 plus or minus 1 inch, so 24
inches is even within the error bars.
kind of normal distribution for the measuring error, the probability of
that is clearly less than 0.5.
To call something two feet out when the probability of that is actually clearly less than 0.5 is bad coaching.
Of course it would be correct to say "It *might* have been more than two
feet out, given measurement errors" especially when you'd add "but most likely it wasn't".
Anyway, the point is that yourThat would be an analogy of saying in backgammon "that play is a big
serve is way wide."
error", leaving out further specification that you don't have any good evidence for.
The player replies, "That's the saddest thing I've ever heard you say. Maybe you should write to the National Institute of Standards and Technology and tell them that their definition of a foot is wrong."A whopper is defined as a range upwards from a threshold value, unlike a foot.
It's not so much sad as it's a misunderstanding of terminology and how
to use it correctly.
Stick's example rollout with few trials, pointed out the problem of
calling a bg play a whopper just because it's possible within a 95% confidence range.
The tennis analogy would be using a device that is plus or minus 30 inch
and then calling something two feet out when it was actually inside the
lines by 5.5 inch.
"Don't you know what it means to use approximate language?"Saying that a play is a whopper when it most likely isn't, is more
inaccurate than it is approximate.
Etc. Do you think the coach's behavior is unreasonable here?Yes. The coach should have clarified himself with "it's almost a
whopper" or "it could be a whopper", instead of insisting that it *is* a whopper.
--
Zorba
On 1/17/2022 4:01 PM, Zorba wrote:
A whopper is defined as a range upwards from a threshold value, unlike a foot.If this were true, I could see your point. Checking bkgm.com again, I
see that it agrees with you, but it doesn't match how I've seen the term used.
The historical post I cited credited Magriel with defining a whopper to
be 0.10, with "whopper with cheese" being somewhat more than 0.10 and a "double whopper" being .20. I've never seen anyone try to define
"cheese" precisely but people still use the term "whopper with cheese"
in this way and they use "double whopper" as well, which suggests to me
that they're still following Magriel. This usage isn't consistent with "whopper" meaning *anything* greater than 0.10---people don't refer to
.20 errors as whoppers. On the other hand, it *is* consistent with the
words having *approximate* meanings---a whopper is approximately 0.10,
a double whopper is approximately 0.20, and a whopper with cheese is something in the middle that is far enough from 0.10 and 0.20 to make
someone want to distinguish it from 0.10 and from 0.20.
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 285 |
Nodes: | 16 (2 / 14) |
Uptime: | 69:33:48 |
Calls: | 6,488 |
Calls today: | 1 |
Files: | 12,096 |
Messages: | 5,275,382 |