How To Record a Game


This page is in need of attention.

Usual method

Go games have been recorded for centuries using paper and pen. Today, a game record may also be created using an SGF editor, but the manual approach is still the most convenient in many cases.

First, obtain a game record form. Mark the handicap stones, if any, by simply drawing a black-filled circle for black stones, and an empty circle for white stones.

The number of each move is simply written on the location where it was played. Black moves are entered with a black pen and White moves with a different color such as red. Two-color pens are sold for this purpose (for example, [ext] , or you may just use two pens, taped together if you please. The point of using two colors is simply to make the game record easier to read. Use a pencil if you prefer to make it easier to erase mistakes. In any case, the first black stone will be a black "1", the first white stone a red "2", etc. Whether or not you use two colors, you may wish to circle the moves of one color, normally white, to distinguish them.

If a move is made on a point where an earlier move was made, e.g. in a ko, you write the number of the move (with its correct colour) underneath the diagram followed by the number of the previous move. For example, if Black 49 takes a ko on the point where Black 37 was played, you write "49 at 37" or even "49 takes ko at 37". If the location was previously occupied by a handicap stone, use the "49 at left of 22" notation, where 22 is a move played on the point to the right. It's possible, though less common, to use board coordinates to indicate where a move was played.

If the point where a ko is being taken is obvious, it suffices to just write "49 takes ko". A ko or other connection might be "49 connects".

If the move is a pass you just write "Black 183 pass" underneath.

It's very easy to make a mistake while taking a game record, and difficult to recover. The most common error is forgetting to increment the move number. This can be avoided by quickly checking mentally each time you record a move that the move number is odd if it's a Black move and even if it's a White move (in a handicap game, of course, White's moves will be odd-numbered).

See also

Comments and discussion


axd: the suggested approach has several flaws that could be corrected. A better approach should deal with the fact that the recorder cannot overwrite or erase anything.

  • colour should not be used, because an intersection can change colour
  • black moves should always be odd numbered. this will help the recorder notice sequencing mistakes
  • handicap stones should be marked with odd move numbers incremented normally, while white should record intermediate passes. No circles should be used for this.
  • if needed, circles could be used to mark all but the one last move so that the last move number can always be found back if needed
  • moves on top of each other are recorded not on the grid but in a separate line started with the intersection coordinates, a colon and then a list of successive move numbers for that intersection (remember that the first occupation of that intersection is still recorded in the grid) (eg K13: 22,65,126, ...)
  • similarly, there should be a line outside the graph to record passes (eg. PASSES: 211,245, ...)

A similar line could be devoted to noting moves that took time (see remark of Bill further).

Alternative recording methods

larrye2000?: I am working on a program that uses a webcam to watch a go game and automatically takes pictures at 10 frames a second. The program then automatically analyzes the images and creates a game record that you can edit and then export to SGF. Info is here with a link to a video. [ext] (invalid link)

maruseru: Has anyone ever used a digital camera, or a mobile phone with a camera, to record moves? A picture every minute or so (depending on playing speed, of course) could later be used to reconstruct the game (e.g., transfer the images to a computer and use an SGF editor). With the new Java-enabled phones it might even be possible to program it to make a picture regularly, so all you need to do is position the phone and start the recording program (but I don't know how far the integration between phone and Java goes).

Ileden: To take it further, you could even write a program to automatically analyze the pictures and create an SGF file from it. I'm guessing it shouldn't even be too hard - after all, the board is a pretty simple shape and the stones have very good contrast.

At least the analysis should be easy if the camera is positioned before play on a tripod or such. Also, the signal to take the picture could be integrated with a game clock. Or indeed camera (or why not a webcam) could take the a picture every second. The analysis program could then determine from the pictures which of them has the next move by checking if there is a difference on the game state. This might even be used to eliminate errors in picture analysis: if a state cannot follow from the previous state, then there must be a problem with this (or the previous) state, and so on.

Phelan:Two interesting projects in this area are GoCam and Chris Ball's [ext] image2sgf.

thombreSoft?: The iOS app Baduk Cap has automatic move recognition and allows to take pictures at pre-defined time intervals.

joshual000: At AGA Go Congress there was a guy doing a variation on this. Whenever he had a game reviewed by a pro he'd take a picture (digital camera) of alternate positions/diagrams.

tapir: I used a camera to take a picture of the final position and once during a game as a help for reconstructing the game. Taking a picture every minute would be quite annoying and not much of a help (messing with dozens of pictures isn't much easier than reconstructing the game). The idea should be to help you in reconstructing it, not technical overkill. Similar to translator notes or so...

murdo?: I have used a sheet of peel off numbered stickers, black and white for odd and even numbers. These eliminate the common error of writing a number out of sequence. I think they were available from the British Go Association shop.


Harleqin: Actually, it is quite usual that also in handicap games the white moves are counted beginning with 2. It always feels very weird when the white moves are odd-numbered.

ChadMiller notes: If the handicap stones are not specially treated in the beginning, then the ko notation "49 at left of 22" special case need not occur. I advocate numbering the handicap stones normally, just as if white had passed the first N times.

Andrew Grant: Of course, if you're just recording games for your own benefit you can use whatever system you like. But we don't really need multiple methods of recording games, and if you intend your game records to be understandable by others it is better to use the existing system as it is used by millions throughout the Far East, in which handicap stones are not numbered.

uxs: I've been trying to record a game twice now: the first I lost count somewhere around the 80th move, and the second I was able to complete. It was 200 moves or so. While it's nice to have a record, I feel that I had to pay a lot of attention to it, which was probably detrimental to my game. Is the benefit of having records (which enables you to do reviews afterwards), worth the downside of not being able to pay 100% attention to the game ?

Niklaus: Right after buying a palm, I used it to record some tournament games. I soon stopped doing it, because it was a distraction. Nowadays, I bring it along to the tournament, and if I want a record of a game, I just do it after the end of the game. At the latest in the endgame my memory gets fuzzy, but usually the moves worth reviewing are early on in the game.

Velobici: Using a Palm to record games causes me to slow down a little and consider my moves more carefully...I have found that it prevents go-blindness, in which I occasionaly dont see an atari due to being too focused on a different move. So, my experience is that recording ones games is beneficial. I am also making it a habit to replay at least part of the games that I play on KGS.

Chris Hayashida: I use my Palm, even during tournament games. I find that it's more useful for study to have the games, even if it might be detrimental to my play. I learn more from having a record of my "serious" games then I do from trying to remember three games at the end of the day. I also think that my memory is faulty -- the parts that I don't remember well, like ko fights and the end game, are also parts where I'm making big mistakes. I usually can remember the opening and the beginning of the middlegame, but I don't feel I have as many problems there. I guess I think that it's more important to have a record of my games for study, instead of having my *best* game. That being said, I think the Palm is the easiest way to record, without having to remember what number you are on, who's move you recorded, etc. I was able to record fairly easily, even in 30-second byouyomi.

Bill: Back when I was playing more or less seriously, I found it very helpful to record my games in real time. I also made a note of which plays took me more than one minute to play. (That's an idea of Botvinnik's. Plays that take a relatively long time indicate positions that you find difficult and thus suggest where to focus your study.)

ThaddeusOlczyk: I used to do it quite simply. I drew a board with ruler. Then I went to the departmental ( college ) photocopier and copied it. Still have about a have dozen of these sheets in a file folder.

I think it has some advantages to follow the chess practice of recording a players move before thinking and recording your move before playing. It did sometimes annoy other people.

Of course you use two different color inks ( usually blue/black/blue-black and red ).

What I'm suprised is no one mentioned something that would often happen to me. I would write down one move sifted by one, then use it to orient my self by it. About twenty moves later, I would realise my mistake and started fixing things. Big mess. The best thing would be to just continue on another sheet. Usually I would be able to piece together the game from the sheets.

Robert Pauli: Recording should be done after one's move -- especially if the record is a diagram. It's like trying out a move: never!

deft: I always note moves that are played in the same place as another (due to ko or captured stones) by _first_ writing the place, then "..." followed by other moves in that location separated by "," . This is much more convenient in long ko fights, since all the moves that go into the ko are recorded in the same place.

Tamsin: If you have a pen of only one colour, then to distinguish black and white you could write White's moves in Chinese numerals.

Malcolm: Three remarks.

1) When using a single pen with just one colour, simply omitting the circle around black moves is enough to distinguish them from white moves.

2) I tend to use a notebook whose paper already has a grid drawn on it in light blue - these are very easy to get hold of. Then I draw out the board on the paper by hand by putting dots on the star points, and drawing a border freehand, just outside the edge of the 19x19 area. This doesn't take very long to do.

3) When recording the game I use a variant on the method described above, that saves space.

tapir: One Berlin enterprise sponsoring go events hands out kifu forms which come together with a transparent page. The idea is you add the moves 1-100 on the kifu itself, 101++ on the transparent page. That makes replaying the game later a bit easier.

other observations, anecdotes, ...

Chris Hayashida: I have to record games when I play teaching games with beginners, since I find it's hard to remember their moves. Sometimes I don't understand the reasoning behind their moves. Anyway, last night, I was caught! The other player said, "Hey, you recorded my move before I played!" (I was right, though.) He thought it was funny, though.

AshleyF: Related funny thing: A friend of mine and I have a habbit of each picking up the stylus and recording our own moves on the one device. Once while playing, he was staring at the Pocket PC instead of the real board while thinking and then slapped a stone down on the screen! Oops!

sle?: Once while recording the game on a smartphone app while playing, I touched the real board like a touchscreen and for a split second was confused why no stone appeared.

Imagist: On a similar note, I was playing a (very) informal game with a friend, and was eating potato chips as I played. The game became interesting, and at one point, while I was concentrating on a particularly difficult sequence, I put a stone in my mouth, thinking it was a potato chip.

axd: Once, I was so concentrated that I found myself grabbing for a stone in my cup of (hot) coffee...

ludwig?: I had the same experience. I once dipped my hand into a cup of tea instead of my bowl of stones. Twice in the same game, no less...

KBunny: For a while I got to play with a group on a daily basis; once a month someone would grab for a stone, and fingers into coffee. Weekly someone would put out their cigarette in their coffee. Felt lucky I got to play daily.

How To Record a Game last edited by on December 9, 2019 - 22:03
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Login / Prefs
Sensei's Library