Perhaps you could check a database to see how often each of these stones dies in actual play, and use that for a weight? Not sure how much more accurate it would make the assessment, but it might show the pro sentiment toward each stone.Bill Spight wrote:
Here is my guesstimate, based upon my study of influence. I estimate the value of a 4-4 stone on the empty board as 14.5, but that is probably a bit high, because it assumes that the stone is immortal.
* snip *
Black might pincer instead, but this follow-up is normal. Evaluating this position is trickier, because plainly the assumption that the White stone is immortal is wrong.
How many points do mistakes lose?
-
Mef
- Lives in sente
- Posts: 852
- Joined: Fri Apr 23, 2010 8:34 am
- Rank: KGS [-]
- GD Posts: 428
- Location: Central Coast
- Has thanked: 201 times
- Been thanked: 333 times
Re: How many points do mistakes lose?
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: How many points do mistakes lose?
When I first noticed, a few years ago, that my influence estimates were surprisingly accurate even though they assumed that stones in the opening were immortal, I did compare pro game records at move 30 vs. final positions, and found that, indeed, nearly all of the live stones on the board at move 30 were still alive at the end of the game. OC, alive does not mean immortal. But based upon that, I expect that the White stone will live more often than not.Mef wrote:Perhaps you could check a database to see how often each of these stones dies in actual play, and use that for a weight? Not sure how much more accurate it would make the assessment, but it might show the pro sentiment toward each stone.Bill Spight wrote:
Here is my guesstimate, based upon my study of influence. I estimate the value of a 4-4 stone on the empty board as 14.5, but that is probably a bit high, because it assumes that the stone is immortal.
* snip *
Black might pincer instead, but this follow-up is normal. Evaluating this position is trickier, because plainly the assumption that the White stone is immortal is wrong.
So let's suppose that we decide, based on the statistics, that the value of 13 pts. is an over estimate. That would mean that pros are often making the slide too early. The trouble then is that we are using incorrect pro play to estimate the value of the slide. Tilt!
BTW, it may be that they are playing the slide too early. Centuries ago a popular opening was to make the first three plays in the same corner: 3-4, 5-3 approach, pincer. After a while the pincer fell out of use, and then a while later the immediate approach fell out of use. We may see the same sort of progression. Pros have started to tenuki in response to the slide. Maybe the slide will fall into disuse as well.
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.
-
Uberdude
- Judan
- Posts: 6727
- Joined: Thu Nov 24, 2011 11:35 am
- Rank: UK 4 dan
- GD Posts: 0
- KGS: Uberdude 4d
- OGS: Uberdude 7d
- Location: Cambridge, UK
- Has thanked: 436 times
- Been thanked: 3718 times
Re: How many points do mistakes lose?
Interesting that you think it is still essentially even, I said ~3, and SmoothOper >6. Also 3-3 a mistake is a bit controversial as it is the most common pro continuation, but jump out does have a better win rate in GoGoD so maybe you are onto somethingBill Spight wrote: I think that we can agree that a wedge by Black is urgent now. I like.
I also think that this position is simple enough that my influence estimates are pretty good. Given White to play and komi, I think that this position is about even.
Not that White did not block from the wrong side, but I expect thatwas a mistake. Better jump out, or possibly make a double kakari.
----
BTW, if you are going to add and subtract mistakes, be sure to use miai counting. Things do not add up with deiri counting.
Building that big wall joseki certainly has a nice consistent feel with double 4-4s.P.S What do you mean deiri doesn't add up? All my swings are for single moves. I'm treating them like "how much would you have to increase/decrease komi by to feel the game is still even" and that adds up.
-
Uberdude
- Judan
- Posts: 6727
- Joined: Thu Nov 24, 2011 11:35 am
- Rank: UK 4 dan
- GD Posts: 0
- KGS: Uberdude 4d
- OGS: Uberdude 7d
- Location: Cambridge, UK
- Has thanked: 436 times
- Been thanked: 3718 times
Re: How many points do mistakes lose?
To all the people saying it is hard to put a value on mistakes, I agree, and they aren't supposed to be accurate. But I like trying hard things. Also a large part of the interest for me was to see how close the sum of the mistakes was to the final score: when you play a stronger player and they just beat you by miles without killing anything I often wonder "Where did I lose all those points?" and this was an attempt to look at that a bit more qualitatively. 35 was pleasingly close to 40.
-
Mef
- Lives in sente
- Posts: 852
- Joined: Fri Apr 23, 2010 8:34 am
- Rank: KGS [-]
- GD Posts: 428
- Location: Central Coast
- Has thanked: 201 times
- Been thanked: 333 times
Re: How many points do mistakes lose?
Yes, this is what I was somewhat hinting at with my last sentence. Using statistics from pro games to weight analysis may simply tell us how much pros currently think the move is worth, rather than how much it is truly worth.Bill Spight wrote:The trouble then is that w e are using incorrect pro play to estimate the value of the slide. Tilt!
The immortal stone assumption you make seems elegant to me. Especially at a professional level, any stone that hasn't been "killed" out right by move 30 would probably only die in a trade (presumably giving equal value to what the live stone was worth). The assumption will err occasionally, however barring a blunder in the game it should err in a way that cancels itself (e.g. two equal value immortal stones dying). Certainly for over the board evaluation, it seems like a useful shortcut.
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: How many points do mistakes lose?
You know, I look at that position and, even though I know that Black has one more stone, so it will look good for Black even if White is OK, I would get a sinking feeling as White.Uberdude wrote:Interesting that you think it is still essentially even, I said ~3, and SmoothOper >6. Also 3-3 a mistake is a bit controversial as it is the most common pro continuation, but jump out does have a better win rate in GoGoD so maybe you are onto somethingBill Spight wrote: I think that we can agree that a wedge by Black is urgent now. I like
I also think that this position is simple enough that my influence estimates are pretty good. Given White to play and komi, I think that this position is about even.
Not that White did not block from the wrong side, but I expect that
----
BTW, if you are going to add and subtract mistakes, be sure to use miai counting. Things do not add up with deiri counting.
IIUC, you would say that this play is a 1 pt. error, because the local result is B +1, while Black could have played to a local result of B +2. This assumes that the play below is the best play on the board.P.S What do you mean deiri doesn't add up? All my swings are for single moves. I'm treating them like "how much would you have to increase/decrease komi by to feel the game is still even" and that adds up.
That's what I meant by using miai counting.
Using deiri counting one would say that
in the first diagram is a 2 pt. move, while in the second diagram is a 4 pt. move, so in the first diagram is a 2 pt. error.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.
-
Fedya
- Lives in gote
- Posts: 603
- Joined: Tue Apr 20, 2010 8:21 pm
- Rank: 6-7k KGS
- GD Posts: 0
- Has thanked: 43 times
- Been thanked: 139 times
Re: How many points do mistakes lose?
I wish I could play a game where my biggest mistake was eitherBill Spight wrote: I think that we can agree that a wedge by Black is urgent now. I like
I also think that this position is simple enough that my influence estimates are pretty good. Given White to play and komi, I think that this position is about even.
Not that White did not block from the wrong side, but I expect that
or 
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: How many points do mistakes lose?
Me, too.Fedya wrote:I wish I could play a game where my biggest mistake was eitherBill Spight wrote: I think that we can agree that a wedge by Black is urgent now. I like
I also think that this position is simple enough that my influence estimates are pretty good. Given White to play and komi, I think that this position is about even.
Not that White did not block from the wrong side, but I expect that
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.
-
ez4u
- Oza
- Posts: 2414
- Joined: Wed Feb 23, 2011 10:15 pm
- Rank: Jp 6 dan
- GD Posts: 0
- KGS: ez4u
- Location: Tokyo, Japan
- Has thanked: 2351 times
- Been thanked: 1332 times
Re: How many points do mistakes lose?
Just keep in mind the list of 'duffers' who made the mistake of jumping into the 3-3 since it first appeared in 1990.Uberdude wrote:Interesting that you think it is still essentially even, I said ~3, and SmoothOper >6. Also 3-3 a mistake is a bit controversial as it is the most common pro continuation, but jump out does have a better win rate in GoGoD so maybe you are onto somethingBill Spight wrote: I think that we can agree that a wedge by Black is urgent now. I like
I also think that this position is simple enough that my influence estimates are pretty good. Given White to play and komi, I think that this position is about even.
Not that White did not block from the wrong side, but I expect that
...
...
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21
-
John Fairbairn
- Oza
- Posts: 3724
- Joined: Wed Apr 21, 2010 3:09 am
- Has thanked: 20 times
- Been thanked: 4672 times
Re: How many points do mistakes lose?
I lack both the knowledge and the intuition to comment on this, but, even on the basis that a little knowledge is a dangerous thing, I'll venture to query it and be the Aunt Sally for the forum.BTW, if you are going to add and subtract mistakes, be sure to use miai counting. Things do not add up with deiri counting
Miai and deiri refer to Japanese accounting systems. In go, boundary-play miai has nothing to do with alternative points, but reflects the accounting idea of just 'look and compare' and is essentially a form of single-entry bookkeeping. It counts the profit and loss in a transaction. Like all single-entry systems, if applied to a whole business or a tax-year, or the current count of a whole game, it is prone to errors that are not caught (because entered only once) and it gives management, or you the player, poor information with which to see the bigger picture. It is the accounting of the school playground, where kids swap marbles for toffees. It satisfies an immediate desire to assess the value of a transaction (in go terms, it measures the urgency of a situation) but is no way to run a life or a business. If you do run a business in this crude way, say as a self-employed person at home, your end-of-your figures may fortuitously work out correctly, but they may not and you may get an unexpected tax bill.
De (out) and iri (in) refer to double-entry accounting. British and American d-e systems vary a little, I know, and the Japanese system has its own nuances, too, but I believe they are all based on the notion that you enter any transaction into two ledgers (or columns), one for outgoings and one for incomings so that when you total up one ledger the result should tally with the total in the other. The basic idea is that you then have two sets of reconcilable data, so that you have a stronger chance of picking up errors or frauds. Even at this basic level I see haze, but since people go on long courses to learn d-e bookkeeping, I suppose I should expect that.
But d-e accounting apparently also has another important function. It provides a vast array of useful information for management. The basis of this idea, as I understand it, is that single-entry (miai) measures Gain = Profit - Loss (what's in front of your nose) whereas double-entry (deiri) measures Equity = Assets - Liabilities (big picture). The essential difference between the two is that miai implies paying in cash immediately for everything to get an immediate (and possibly negative) gain, but deiri means you can hold off on actual payment and count the long-term value of an asset without cashing it at once. You can also play off one asset/liability against another.
Still, assuming I'm somewhere near the truth with that, in theory, if no counting mistakes are actually made and even if one person is driving through a fog and the other through the sunshine, the two systems should at the end of the year/game, come to the same destination/result. In other words they should both "add up". Enlightenement, please!
On a separate tack, talk of miai amd deiri in go always ends up bogged down about which system is meant. Whilst this in part is because of uncertainty over what each system means, I think the ambiguity of certain terms like 'point' doesn't help. Is there not scope to define new units? E.g. deiri points are dits and miai points are mits?
-
tapir
- Lives in sente
- Posts: 774
- Joined: Fri Apr 23, 2010 5:52 pm
- GD Posts: 0
- Has thanked: 137 times
- Been thanked: 155 times
- Contact:
Re: How many points do mistakes lose?
I doubt any professional teacher would declare takemiya-style framework building a mistake - they will tell you that you have to now have to use your influence efficiently and maybe rarely that this requires strength and that you might want to try another more balanced style. but a mistake, I very much doubt it. There is anyway an abundance of (to the professional) obvious mistakes in the usual amateur game so why should they comment on perfectly fine, maybe slightly inferior or more difficult moves? Especially if amateur players of around my level start to build a framework - there will invariably a lot of obvious mistakes later on. Bad invasions, missed focal points, too deep reductions, wasted aji, missed forcing moves etc. ... I never had professionals comment badly on the initial framework building (unless totally misguided) but the teaching opportunity offered by all the follow up mistakes, they never missed to comment on that.Bantari wrote:Consider: hardly anybody today plays like, for example, Takemiya used to... does that mean his moves were weak? Or strong? Can we draw such conclusions?
-
tapir
- Lives in sente
- Posts: 774
- Joined: Fri Apr 23, 2010 5:52 pm
- GD Posts: 0
- Has thanked: 137 times
- Been thanked: 155 times
- Contact:
Re: How many points do mistakes lose?
I guess uberdude and many others here would be quite happy if they could achieve "incorrect pro play" and "incorrect pro positional judgement" level.Mef wrote:Yes, this is what I was somewhat hinting at with my last sentence. Using statistics from pro games to weight analysis may simply tell us how much pros currently think the move is worth, rather than how much it is truly worth.Bill Spight wrote:The trouble then is that w e are using incorrect pro play to estimate the value of the slide. Tilt!
-
Uberdude
- Judan
- Posts: 6727
- Joined: Thu Nov 24, 2011 11:35 am
- Rank: UK 4 dan
- GD Posts: 0
- KGS: Uberdude 4d
- OGS: Uberdude 7d
- Location: Cambridge, UK
- Has thanked: 436 times
- Been thanked: 3718 times
Re: How many points do mistakes lose?
Yes, but playing better than pros is even better!tapir wrote:I guess uberdude and many others here would be quite happy if they could achieve "incorrect pro play" and "incorrect pro positional judgement" level.Mef wrote:Yes, this is what I was somewhat hinting at with my last sentence. Using statistics from pro games to weight analysis may simply tell us how much pros currently think the move is worth, rather than how much it is truly worth.Bill Spight wrote:The trouble then is that w e are using incorrect pro play to estimate the value of the slide. Tilt!
Something that's happened a few times on OGS is I have been playing a game largely the same as a pro game (and fairly sure my opponent is too rather than finding all the moves themselves) but I usually seem to be the pro who lost. So I eventually need to deviate or I will lose. Thus I have to study the pro game intently to find the pro's mistakes; I think I managed to do that in this game! (Of course my opponent helped me by not blocking at q2, and the trade of the corner for the ladder is probably not as good as it ended up in the game as I think white failed to utilise his thickness effectively: o14 suspect, maybe even q14? r10 should be top left? But the q4 push does seem to hurt white's shape and offer an improvement on the pro game).Edit: above game is in progress (but essentially over as I'm 20 points ahead and everything is settled) so the live sgf link breaks, game is at http://www.online-go.com/games/board.php?boardID=394377.
This reminds me of something Guo Juan said: she was at her strongest (beat Sakata when he visted China) after studying the games of Shuei which consisted of trying to find the players' mistakes as this really focuses your attention and makes sure you aren't just a passive observer.
- Attachments
-
- 2006-08-05j.sgf
- (1.53 KiB) Downloaded 546 times
Last edited by Uberdude on Sun Jun 02, 2013 3:25 am, edited 6 times in total.
-
Bill Spight
- Honinbo
- Posts: 10905
- Joined: Wed Apr 21, 2010 1:24 pm
- Has thanked: 3651 times
- Been thanked: 3373 times
Re: How many points do mistakes lose?
John, see http://www.lifein19x19.com/forum/viewto ... 40#p137540 for an example of how go players make mistakes trying to add and subtract deiri values.John Fairbairn wrote:Still, assuming I'm somewhere near the truth with that, in theory, if no counting mistakes are actually made and even if one person is driving through a fog and the other through the sunshine, the two systems should at the end of the year/game, come to the same destination/result. In other words they should both "add up". Enlightenement, please!BTW, if you are going to add and subtract mistakes, be sure to use miai counting. Things do not add up with deiri counting
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.