- gerardTG00.png (2.69 KiB) Viewed 38923 times
At or below temperature 1, however, Black plays the keima to keep sente at the cost of 1 point on the top, but then is able to gain 2 points on the bottom, for a net gain of 1 point.
Thermography
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
How about this one?
At this point, the top and bottom are strictly miai, simple gote that gain 2 points. Each player gains 2 points for net 0 gain. Thus the straight vertical line at a count of -1 down to temperature 1.
At or below temperature 1, however, Black plays the keima to keep sente at the cost of 1 point on the top, but then is able to gain 2 points on the bottom, for a net gain of 1 point.
At this point, the top and bottom are strictly miai, simple gote that gain 2 points. Each player gains 2 points for net 0 gain. Thus the straight vertical line at a count of -1 down to temperature 1.
At or below temperature 1, however, Black plays the keima to keep sente at the cost of 1 point on the top, but then is able to gain 2 points on the bottom, for a net gain of 1 point.
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
It is not that easy but at least I believe I undertood your thermographs in your last posts.
BTW I continue to fail finding an environment in which kosumi is better than both keima and monkey jump. It seems that if this environment exists then after
if 
black should play in the environment and depending on white answer, black chooses "a" or "b" accordingly (but I am not completly sure of that fact!)
BTW I continue to fail finding an environment in which kosumi is better than both keima and monkey jump. It seems that if this environment exists then after
if black should play in the environment and depending on white answer, black chooses "a" or "b" accordingly (but I am not completly sure of that fact!)-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
Concerning the building of the left wall in the position above, is it true that beginning with
is correct (but not the only solution) at all temperatures?
is correct (but not the only solution) at all temperatures?
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
Yeah, me too. Not that I have tried hard. It appears that for the keima and the kosumi to be in the running, the best follow-up to the monkey jump needs to be in the environment. Then it seems like keima and the kosumi come into play only below temperature 1. We can prove stuff about play at that level, so maybe there is a proof that the kosumi cannot be dominant in non-ko environments.Gérard TAILLE wrote:It is not that easy but at least I believe I undertood your thermographs in your last posts.
BTW I continue to fail finding an environment in which kosumi is better than both keima and monkey jump.
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
You surely need to work out the scaffold for it.Gérard TAILLE wrote: Concerning the building of the left wall in the position above, is it true that beginning with
is correct (but not the only solution) at all temperatures?
Edit: Anyway, above temperature 1 the play goes like this.
This leaves a gote at temperature 2. The local count is -3.
Between temperature 2 and temperature 1 Black continues, saving her stones.
The local score is -1.
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
Finally the three moves above seems correct whatever the temperature.
After this basic sequence:
-if temperature is above 1 white continue with "a" otherwise white play "b".
-between temperature 1 and 2, after white "a" black answers at "b"
-above temperature 2, after white "a" black plays tenuki.
This result is surprisingly quite simple isn't it?
After this basic sequence:
-if temperature is above 1 white continue with "a" otherwise white play "b".
-between temperature 1 and 2, after white "a" black answers at "b"
-above temperature 2, after white "a" black plays tenuki.
This result is surprisingly quite simple isn't it?
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
Just a basic theoritical question.
For analysing a local yose area I often prefer to use an area counting. Assuming there no seki, my question is the following:
is the thermograph in "territory counting" strictly identical to the thermograph part above temperature 1 in "area counting" ?
For analysing a local yose area I often prefer to use an area counting. Assuming there no seki, my question is the following:
is the thermograph in "territory counting" strictly identical to the thermograph part above temperature 1 in "area counting" ?
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
Consider this case.
Let White play first.
gains as much, on average, as the reverse sente, and lets White get the last play (locally) of that size. After 
the local temperature has dropped to 2.
Now let Black play first.
sets a small trap for White. 
loses ⅓ of a point on average, as discussed above. After 
the local temperature has dropped to ⅚.
maintains the local temperature at 2. Below temperature 2 White plays on.
captures and 
.
Let White play first.
gains as much, on average, as the reverse sente, and lets White get the last play (locally) of that size. After the local temperature has dropped to 2.Now let Black play first.
sets a small trap for White. loses ⅓ of a point on average, as discussed above. After the local temperature has dropped to ⅚. maintains the local temperature at 2. Below temperature 2 White plays on. captures and . The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
Except for the count, yes. And for Japanese seki (edit: in the environment) you just have to adjust the count accordingly. There may be some kos or superkos that different rules treat differently, and that will affect the thermograph.Gérard TAILLE wrote:Just a basic theoritical question.
For analysing a local yose area I often prefer to use an area counting. Assuming there no seki, my question is the following:
is the thermograph in "territory counting" strictly identical to the thermograph part above temperature 1 in "area counting" ?
Edit: OIC. You mean if there is a seki or possible seki in the play. In that case a Japanese seki may well alter the thermograph.
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
tenukyBill, when you analysed the sequence above you clearly took into account the resulting ko by using one third of the deiri value of the ko. That's sounds fine for me.
but after this sequence above I do not understand why you never take into account this ko.
Let me try to count it.
white wins the ko:
tenukiand the score for black is -8
black connects the ko and the score for black is -1
I conclude that each move in the ko is worth 2⅓
But we have to be aware that after: black may also defend at "a"
Let's count this new position: black to play
tenukiand the score for black is -1
white to play and the score for black is -6
Eventually the score of this position is -3½
tenukicomparing with this sequence leading to the black score -8 it means that the black defense at "a" is worth (8 - 3½)/2 = 2¼
Seeing this black defense is worth 2¼ which is less than the value 2⅓ of a move in the ko I conclude that black will never defends like this because when black chose to keep sente that meant the temperature was above 2⅓.
Finally when black, above temperature 2⅓ plays 1, the score of black is not -3 but rather -3⅓
If all that is true then the thermogrph corrected is the following
- grille3.png (25.92 KiB) Viewed 40547 times
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
Since ko thermography is complicated and we are just getting started (I am including many of our readers in that.Gérard TAILLE wrote:
Bill, when you analysed the sequence above you clearly took into account the resulting ko by using one third of the deiri value of the ko. That's sounds fine for me.
but after this sequence above I do not understand why you never take into account this ko.
), I avoided discussion of ko fights, except for suggesting the traditional assumption of no ko threats. This position confounds ko with sente, which makes it fairly advanced. Nope. Sente is involved.Let me try to count it.
white wins the ko:
and the score for black is -8
black connects the ko and the score for black is -1
I conclude that each move in the ko is worth 2⅓
Good catch.But we have to be aware that after: black may also defend at "a"
In real life Black might play at 4 instead of connecting at 3, threatening the life of White's group.Let's count this new position: black to play
and the score for black is -1
But assuming no ko threats, White will just take and win the ko. So let fill the ko. With sente, as your indicates. Black has played only one more move than White, unlike with a regular ko.Well done!white to play and the score for black is -6
Eventually the score of this position is -3½
In calculating the count you correctly took sente into account. Edit: Here I missed the miscalculation of the score when White wins the ko. It is -5, not -6.
Note also that each play in the ko gains 2½ points. Playing the ko raises the local temperature.
Edit: That means that each play in the ko gains only 2 points, not 2½ points. Since 2 < 2⅓, Black's connection in the corner reduces the temperature, not raises it. I corrected my mistake in my next note.
That might be so if Black a were gote, but it is sente. You can see that by the fact that 2¼ < 2½.
comparing with this sequence leading to the black score -8 it means that the black defense at "a" is worth (8 - 3½)/2 = 2¼
Edit: Again incorrect, based upon the inaccurate calculation.
Black a threatens to take and win the ko, also in sente.
So, assuming no ko threats, if White takes the ko, play goes like this.
The local score is -6 with one net play by White.
Edit: Correction. The local score is -5. The rest is fine.
achieves the same result without the ko. In a real game White might take the ko to eliminate the possible Black ko threat in this position. But the thermograph is the same.
Last edited by Bill Spight on Thu Oct 08, 2020 3:02 pm, edited 2 times in total.
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
[quote="Bill Spight"]
Oops I do not see how can be sente.
Instead of answering at "a" white can be the first to play in the environment, answering at "a" only after black takes the ko can't she?
That is the basic reason why I think the ko gives white a slight advantage comparing to the analyse without ko.
can be sente.Instead of answering at "a" white can be the first to play in the environment, answering at "a" only after black takes the ko can't she?
That is the basic reason why I think the ko gives white a slight advantage comparing to the analyse without ko.
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
Actually, my characterization ofGérard TAILLE wrote:Bill Spight wrote: Oops I do not see how
Instead of answering at "a" white can be the first to play in the environment, answering at "a" only after black takes the ko doesn't it?
here as sente is not accurate. It was based on a miscount. 
Let's do it right.1
prisonerThis position, assuming no ko threats, has a count of -3, counting one White prisoner which Black has captured.
1
prisonerThe local score is -1. With one net play Black has gained 2 points on average.
Now suppose that White wins the ko in 2 moves.
1
prisonerThe local score is -8. If this line of play is correct, there is a difference of 7 points in 3 plays, for an average gain per play of 2⅓ points. (And the original value of -3 is wrong. It should be -3⅓.)
But suppose that
connects in the top left corner.1
prisonerNow we have a new ko.
1
prisoner, 1 
prisoner The position after
has a score of only -5, not -6. What were we thinking? 
Let Black take and win the ko, with sente.
1
prisoner, 1 prisoner fills the ko with sente. Alternatively, White can play at 4. The result is -1 with one net Black play in either case, OC.That means that each play in the ko (except Black filling it) gains 2 points, not 2½ points.
1
prisoner, 1 prisonerAnd that means that the position after
is worth -3, on average. And that means that gains 2½ points. 1
prisonerAnd that means that the position after
is worth 5½ points, and is indeed gote, gaining 2½ points. And that means that raises the local temperature from 2 to 2½. is the sente, not . Sorry about that.
Anyway, before White takes the ko the local temperature is 2.
raises it to 2½, then brings it back down to 2. It is true that 
can tenuki, but the sente exchange, - , has not altered the count. Only now, Black can take and win the ko with sente.The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Thermography
This time I quite agree with you Bill
Before temperature drops to 4 black plays in sente: when temperature drops between 2 and 2.5 white continue in sente: but here is the point. Can white gain something, even if the following capture of the ko by black is sente ?
From a theorical point of view it seems possible.
Assume white waits until we reach the temperature 2+ε, I mean the temperature of the smallest point above temperature 2.
Then the following sequence will take place:
takes point 2+ε
connects
takes point at temperature 2
As you can see, with this strategy white can gain ε points.
In any case, without any calculation, when seeing the position white can answer by the safe "a" move but, undoubtely white might in certain circumstancies, be confident that black cannot really lose another move by playing at "a" even for a far larger ko.
That means that, from a theoritical point of view you can assume that this possibility allows white to gain say ε points (in practise ε may be equal to zero but sometimes ε may be greater than zero).
Before temperature drops to 4 black plays in sente: when temperature drops between 2 and 2.5 white continue in sente: but here is the point. Can white gain something, even if the following capture of the ko by black is sente ?
From a theorical point of view it seems possible.
Assume white waits until we reach the temperature 2+ε, I mean the temperature of the smallest point above temperature 2.
Then the following sequence will take place:
takes point 2+ε connects takes point at temperature 2As you can see, with this strategy white can gain ε points.
In any case, without any calculation, when seeing the position white can answer by the safe "a" move but, undoubtely white might in certain circumstancies, be confident that black cannot really lose another move by playing at "a" even for a far larger ko.
That means that, from a theoritical point of view you can assume that this possibility allows white to gain say ε points (in practise ε may be equal to zero but sometimes ε may be greater than zero).
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: Thermography
It now appears that the following sequence may be dominant play for White.Gérard TAILLE wrote:This time I quite agree with you Bill
Before temperature drops to 4 black plays in sente:
However, White can possibly reach the same ko position by playing at a when the temperature is above 2 but below 2.5. This sequence does not affect your argument below.
For clarity, here is the position beforewhen temperature drops between 2 and 2.5 white continue in sente: but here is the point. Can white gain something, even if the following capture of the ko by black is sente ?
From a theorical point of view it seems possible.
Assume white waits until we reach the temperature 2+ε, I mean the temperature of the smallest point above temperature 2.
Then the following sequence will take place:
As you can see, with this strategy white can gain ε points.
In any case, without any calculation, when seeing the position white can answer by the safe "a" move but, undoubtely white might in certain circumstancies, be confident that black cannot really lose another move by playing at "a" even for a far larger ko.
That means that, from a theoritical point of view you can assume that this possibility allows white to gain say ε points (in practise ε may be equal to zero but sometimes ε may be greater than zero).
.The number of prisoners is equal.
Because of the absence of ko threats, Black cannot make a ko for the life of the group.
takes the last move before the temperature drops to 2, gaining 2+ε. Our model environment assumes a sufficiently large number of simple gote at temperature 2, followed by a sufficiently large number of simple gote at a slightly lower temperature, etc. Suppose that Black takes the ko in the following sequence.
connects. Black has gained 2 points on average in this exchange. White now plays in the environment, taking a 2 point gote. We estimate White's gain from doing so as 1 point. So our estimate of White's gain, starting with
is 2+ε - 2 + 1 = 1+ε.Note that this estimate is the same as the one if Black took a simple 2 point gote instead of this ko.
Suppose now that
takes a play in the environment and White wins the ko, and then Black plays in the environment again. Both Black plays gain 2 points, because there are plenty of 2 point plays in the environment. gains 2 points, on average.White's expected gain from this sequence is 2+ε - 2 + 2 - 1 = 1+ε, the same as above.
Now let
take the ko but play in the environment, and then wins the ko with sente. elsewhere, connects. White's expected gain from this sequence is 2+ε - 2 + 2 - 1 = 1+ε, still the same.
OC, this is the same as if there were a simple 2 point gote on the board instead of the ko.
----
Now let's use my original model of the environment as a set of simple gote, each gaining gi, such that g0 ≥ g1 ≥ g2 ≥ . . . . Let g1 = 2 and
takes g0. connects. White's expected gain is g0 - 2 + g1 - g2/2 = g0 - g2/2.
elsewhereWhite's expected gain is g0 - g1 + 2 - g2/2 = g0 - g2/2, the same as above.
elsewhere, connects. White's expected gain is g0 - 2 + g1 - g2/2 = g0 - g2/2, the same as above.
All same same.
The result is the same as if the ko were a simple 2 point gote.
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.
At some point, doesn't thinking have to go on?
— Winona Adkins
Visualize whirled peas.
Everything with love. Stay safe.