I've been experimenting on OGS with Capped Fischer timing (called Fischer on OGS) and I'd like to report and discuss some of my experience. Using something reasonable like 5 mins Max Time and, say, 30 secs for the Increment, I find it is still very easy for one of the players to run out of the stored time and end up essentially playing a single-period byoyomi with 30 seconds. Worst of all, this naturally tends to happen when the position is complicated and some careful thinking is required, but which may be difficult to do with just 30 seconds. It looks like the apparent advantage of Fischer timing disappears when it is most critical.

I mostly play on IGS and I use Canadian timing with 25 moves per 10 mins, which gives me 24 seconds per move on average. And I almost never need to think for more than 5 minutes on a single move. So, a roughly equivalent Fischer time setting will be 5 mins Max Time and 24 seconds for the Increment, which is close to what I've been using on OGS. Admittedly, I'm more used to Canadian but I have to say I find it more comfortable since I know I will get my 10 minutes back sooner or later. The problem with Canadian is that sometimes there are several moves left to play but the 10 mins are almost exhausted. However, in my experience this often happens when the game is at some routine stage with plenty of quick moves available.

Are you capping at 5 minutes? That's basically blitz timing in Fischer terms, especially with 30 second increments. Try a longer main time with a shorter increment. 15 + 20 or similar.

This is an unfair comparison. To be fair you would have to use the same max time, i.e. 10 minutes (and increment 24 sec per move), if you are used to 25/10 Canadian.

I have experimented with Fischer 2 min start time, 20 seconds per move, max time 5 min, and felt quite comfortable.

I have played a lot of fischer timed go games (albeit not capped) as I have a physical clock that allows that setting.

I generally prefer the following settings for fast, medium or long games: 10/15, 20/30, 30/40.

I don't like Fischer time. If the increment and initial time are both small then you get the situation you describe. All other situations can lead to what is essentially an untimed game. It therefore seems to serve no useful purpose at all in my opinion. My favorite is definitely Canadian time, but that is probably not surprising as it is the time limits I have been used to at tournaments for most of go-playing life.

_________________
Still officially AGA 5d but I play so irregularly these days that I am probably only 3d or 4d over the board (but hopefully still 5d in terms of knowledge, theory and the ability to contribute).

Canadian time: Add ca. 10 minutes every 25 stones
Fischer time: Add 20 seconds every stone.

Pretty much identical, except the Canadian version is chunked.

What is the advantage of capping the tinking time? In a tournament organisers are happy if they can predict the length of the round with good accuracy. Fisher timing is good in that sense. Capping cuts essentially randow time slices from player's allowance, but gives no guarantee that the slowest game gets finnished faster.

I've seen capped Fischer on turn based servers, e.g. 40 days + 1 day capped at 40. This is more to prevent games being allowed lie fallow for two months between moves though I believe whilst still allowing for more reasonable periods of inactivity not to harm a player too much. This would mostly be because 1 day is a large increment though relative to how fast people could burn through 20 moves.

I think I would consider them medium, long, and really long (=. Those last two time settings could be roughly compared to 45+1:00(5)? The games could easily be 3-4 hours or more!

A reasonable pace, IMX, for amateur games is to average around 15 sec. per move. The game will take about an hour, assuming that the game lasts around 240 moves. For such a game you want the Fischer increment to be less than 15 sec. Like 10 sec. Then the main time would be 10 min. If the game went 300 moves, that would take another 10 min. at 10 sec. per move. Instead of having a frantic endgame for long games, you could add those 10 min. to the main time to get 15 min. main time, 10 sec. increment. If that seems short, you could make that 20 min. main time, 10 sec. increment. A 240 move game would last around 80 min.

I do not think that it is good to think of the Fischer increment like we do byo yomi. Yes, if you have no main time left, then you have to make one move within the Fischer increment, but you should have enough main time to start with so that that should seldom be the case. In the endgame you should be able to make most moves more quickly than the Fischer increment, so you will be adding time with those moves.

So, to make the comparison with a Canadian byo yomi of 25 moves in 10 min., which is an average of 24 sec. per move, a Fischer increment of 15 sec., with 3 min. 45 sec. added to the main time for those moves, would be comparable. An initial setting of 45 min. main time, 15 sec. increment, might be good. A 240 move game would last around 2 hr. 30 min.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I'm only talking about casual games, not tournaments. I want the pace to be more or less the same during the game, so I tend not to use any main time at all. I like to keep the max time below 5 mins since I don't want anyone taking longer than 5 mins to make a move. In my experience, increasing max time to 10 mins doesn't help in any case. I've seen people lose their 10 minutes just as easily. Setting the increment to a small number like 10 secs is dangerous and unfair. All too often one player will end up having to play 10 sec blitz while the other has several minutes to spare. It's not a very good game for both. But even if the increment is, say, 30 secs, it is still quite short and not much better than byo-yomi.

In the end, it's all down to proper time management. I think one can learn to use Fischer timing successfully with enough experience. One has to learn to pace oneself correctly when the position is complicated and stop relying on the increment. Still, Canadian timing feels a bit more forgiving than Fischer and easier to learn to use properly. Perhaps the reason Fischer is more difficult is that it's hard to know how many moves one will need to play to resolve the complications, whereas with Canadian timing one has a certain number of moves left which sets the pace.


Have you ever come across the situation I described?


I think your approach to using Fischer timing is very different from mine. I want to set a certain average pace, yet allow some extra thinking time if the position gets complicated. With your settings, the game will be very similar to 20 min main time followed by 10 sec byo-yomi (or something quite similar to it). I'm curious, have you actually tried playing with these time settings? What was your experience. Note that you can't set main time larger than the max time on OGS.

I think that this comes down to lack of experience with the Fischer set-up that you have been experimenting with. If you are really reduced to something like a single byo-yomi situation it is because you are not accumulating any extra seconds under Fischer. In order for this to happen, you (or your opponent) must be using more time per move than is available under Canadian 10/25! In other words, if you were using the same average time per play under both systems, you can't be using more than 24 seconds since that is all that the Canadian set-up delivers to you. At 24 seconds per play, you will rebuild your Fischer 'stock' at a rate of 150 seconds (2 1/2 minutes) per 25 plays. But this is the ideal case for Canadian - you are using all the available time delivered by the system (much more typical is to 'spill' unused time on every 25th move). This is unlikely because it translates to 48 minutes for each player in a 240-play game. If this were really the pace that you played at you could just use uncapped Fischer at 24 seconds per play and have a more flexible use of time than under Canadian. A much more likely scenario is that you are playing quite quickly normally (hence keeping a big reserve under Canadian and giving up lots of available time under Fischer since the cap is full) until you get into trouble. Once in trouble you start to use all your 30 seconds for each move under Fischer and never regain a cushion. Under Canadian you think you have more flexibility (especially when that next 600 seconds comes crashing in) but in reality the system is delivering less total time to you (24 seconds average versus 30). Hence it is not the system that is the problem but rather lack of sufficient practice with it, IMHO.

An easy test is to increase the cap to 10 minutes (to match the same max as under Canadian) and use a 24 second increment. Note that now both systems are delivering the same amount of time (600 seconds every 25 plays). I think you can only be worse off under Fischer if you are actually using more total time than under Canadian. What do you think?

_________________
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21

What that indicates is that the players do not have enough main time.



Not really. Even if your max thinking time per move is 5 minutes, that does not mean that you don't take more than 24 sec. on other moves. Once you have used up that 5 minutes main time, unless it is late in the game, it will be like Japanese byo-yomi, not Canadian byo-yomi. (As you indicate above.)



Yes, you feel comfortable having that extra time. It gives you flexibility. The Fischer increment will not give you that. It is more flexible than Japanese byo-yomi, but not as flexible as Canadian byo-yomi. To get flexibility you have to have enough main time. Try 20 min. main time, with a 20 sec. increment. You may find that it yields a comfortable pace without making the game too long.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

The best way to think of Fischer time is as an improved version of absolute timing. You still have to manage your main time throughout the game, but you have the benefit of bonus time to play the game out and the possibility of banking time for a few longer thinking periods during the late middlegame / endgame.

Everyone seems to want to use Fischer for some novel blitz settings, but this is much more appropriate in chess than go.

Or as an improved version of Japanese byo-yomi.



Towards the end of the game you should reach a point where you are adding time with most moves. If you set the increment to the average time per move, that is the case at the start of the game. That might be appropriate for blitz, but not for regular games. That is why you want to set the Fischer increment below the average time per move.

How far below? Well, we run into the paradox of byo-yomi. I remember when I first played with byo-yomi. It was 30 sec. per move. Although I averaged 15 sec. per move, having to make a play in 30 sec. was nerve-wracking. And that was so, even though towards the end of the game I probably averaged something like 5 sec. per move.

Let's compare 30 sec. byo-yomi with a Fischer increment of 10 sec. Suppose that I have run out of main time. Well, I might have to scramble to make a play in 10 sec., but on average I am going to increase my main time to 5 sec. After only four moves I will have 20 sec. on average on my clock, and 30 sec. to make a move. OC, if I take almost the whole time I will have to start building my time up again. But I should be at a stage of the game where I rarely have to think for 30 sec. Personally, I play the endgame very quickly, but for most people my guess is that after move 200 would be OK. The idea, then, would be to have enough main time so that I expect to have 30 sec. left at move 200.

Suppose that I am used to Canadian byo-yomi at 25 moves in 10 min. That's 100 moves in 40 min., averaging 24 sec. per move. If the Fischer increment is 15 sec., then I will have added 25 min., so I would need to start with a main time of 15 min., plus the 30 sec. I want to have left. If the Fischer increment is 10 sec., then I will have added 16 min. 40 sec., so I would need to start with a main time of 23 min. 20 sec., plus 30 sec. Fischer timing requires a good bit of main time.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.