Are there caps?
2 Likes
I think the cap goes against the spirit of fisher.
It encourages players to have to spend time when they reach the limit, rather than playing faster to save time. (But the phenomenon like byo-yomi will be more serious, so okay…)
However, if there is no upper limit, then there may be the evader problem I mentioned earlier🤔.
1 Like
To my taste +10s feel bit blitzy (let say I used my main and have to urge a bit) I would gladly have a +15s or even a +20s to still get enough time to have some survival reading. On a side note not everyone has a good enough connection to play at this speed, and considering the full time setting being the slowest automatch it would be great to stay accessible for those who don’t have the best internet connection.
It’s not because a majority of players will play quickly their “full” that it’s appropriate to push everyone to be a bit hasty (which is still the case with 20s in my opinion even if I know there is the main time to spend and capitalization). Full should follow a kind of relax comfy speed in which we allow some real thinking challenge but with respect to the partner to not get too lost in the head, I mean like referring to the many games I played without clocks at all.
At 10mn+10s you surely will sacrifice a big part of interesting reading and focus on the game. But I guess a lot of players do enjoy it, we just have different ways to play a full game. Now we will have rapid and blitz for those who can’t slow down a bit.
For Blitz and rapid I guess your proposition is fine. (I would maybe suggest a +8 in rapid instead of a +5?)
Finally I was a bit confused by OGS fischer time setting which differs from IRL tournaments. In these the main time simply decreases and only then it is used, you go into a fischer to let you finish the game, speeding the players to finish. On OGS the fischer applies from the start instead and that’s quite different!
I don’t think OGS currently has the ability to run fischer without some kind of cap.
I agree. Hopefully he implements the functionally infinite maximum cap of 4 hrs 
2 Likes
What’s the sprit of fischer?
It’s not a trick question, I just want to understand. From what I assume fischer time was introduced as an alternative to sudden death, so that chess player could still finish the game “by hand” even if he’s low on time. I have no idea whether my assumption is correct or wrong.
1 Like
Imo the core of Fischer is very simple. You get some time. After each move, you get some more time
In correspondence, it can make sense to cap how much total time one may have at any given time (OGS is a bit unusual having the option for Live games, but I’m all for options as long as we have the functionally infinite options available). To facilitate commentary and live viewing, it can make sense to cap how much time may be spent on a single move (OGS does not have this option). But both of these are additions to Fischer that make sense in some circumstances, but aren’t the core of what makes it work
Sounds like byo-yomi to me. I guess you should’ve clarified that time could be accumulated to make a distinction between the two.
I was trying to find out how chess players use Fischer time increment
Does what they do go against spirit of fischer? If yes, then why do we need to be such purists as go players?
If no, can’t we do the same and why it isn’t even considered?
I applied the EGD Adjusted Time (TA) formulas to calculate predicted maximum total game time for the time settings for which you listed statistics, ordering them by your observed p99 (i.e. soft max) total game time (listing rounded minutes).
| Time setting | p50 | p90 | p99 | TA |
|---|---|---|---|---|
| Blitz | ||||
| 30s+3s | 10 | 14 | 17 | 16 |
| 30s+5x10 | 11 | 17 | 22 | 16 |
| Rapid | ||||
| 5m+5s | 21 | 31 | 35 | 35 |
| 5m+5x20s | 16 | 27 | 38 | 40 |
| 10m+5x20s | 20 | 32 | 44 | 50 |
| 5m+5x30s | 22 | 34 | 50 | 55 |
| Full | ||||
| 10m+5x30s | 23 | 39 | 54 | 65 |
| 20m+5x30 | 28 | 49 | 67 | 85 |
| 10m+10s | 26 | 52 | 69 | 70 |
| 20m+10s | 37 | 64 | 85 | 90 |
(hoping I didn’t make a mistake there)
(Edit: added 10m+5x20s later on)
From that I note that
- The TA formula gives a pretty good estimate for Fischer p99 total game time. With byoyomi the TA formula is not as good a fit with blitz and full time settings. I suppose that is because byoyomi time gets progressively more wasted as the basic time and the byoyomi period get longer, so the total game time is not increasing linearly with the byoyomi period (while it does with Fischer time). In hindsight that does make sense, but it means that my previous assumption that overtime usage is a linear function of byoyomi time does not hold (and perhaps the TA formula should be modified to account for that. Although I wonder if TA might still be good to predict a more extreme maximum game time like p99.5 or p99.9).
- In practice, full 10m+10s and full 20m+5x30 are pretty much equivalent. My proposals of using either full 20m+10s or 10+5x30s were not good ideas. Your tentative choice seems pretty much optimal.
- With rapid it seems a bit harder to pick a byoyomi setting that fits well to 5m+5s Fischer. If you prioritise maximum total game time (p99 and TA), my proposal of using 5m+5x20s seems to be a better fit, but if you prioritise typical total game time (p50 and p90) and move pacing 5m+5x30s seems to be a better fit. I suppose players should pick Fischer rapid when they have a hard limit of ~40 minutes to finish the game, while they should pick byoyomi rapid if they want a somewhat more relaxed rapid game with a reduced chance of a time-out, accepting that the game might take an hour to finish (usually it takes less than half an hour to finish a rapid game).
Do you perhaps also have statistics for byoyomi rapid 10m+5x20s? I don’t think those settings have been mentioned yet in this thread, so perhaps it’s not very popular, but it might be a nice fit between full 20m+5x30s and blitz 30s+5x10s (at least I now assume it would be a better fit than rapid 5m+5x20s that I proposed earlier).
What are you considering to use as max values for Fischer time bank for blitz/rapid/full?
2 Likes
Yeah, I’m not really sure why that’s not the case, but I learned that during this process too.
Honestly with Rapid, 5+5x30s is just a very common setting that folks seem to like, that was the prioritization there. Aesthetically, and perhaps even speed wise, I think 5+5x20s seems like a great pick too, I just want to give folks what they want though and stats wise I think folks want that 5+5x30s option. I picked the Byo-Yomi settings first and then used those expected values to try and pick decent corresponding Fischer settings.
I would love to provide more granularity with speed options like Lichess has. 5+5x20s would fit great between 30s+5x10s and 5+5x30s and it wouldn’t surprise me if it turned out to be a popular choice given a big enough player pool that folks could find matches regardless of their choice, I just don’t think we’re there yet.
| Settings | Game duration | Moves | Avg. Move time | Samples | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Avg | 50th | 90th | 99th | Avg | 50th | 90th | 99th | Avg | 50th | 90th | 99th | ||
| 10m + 5x20s | 23:16 | 20:09 | 31:46 | 44:05 | 206 | 210 | 284 | 327 | 6.65s | 5.84 | 9.07 | 12.46 | 1747 |
It seems like the general desire is high enough so that you don’t bump into it in practice, so maybe something like 10m for blitz, 30m for rapid, 120m for full or something?
3 Likes
So long as you’re not fully disconnecting, latency to the servers doesn’t matter anymore, you have your full move time. It’s been like that for a few years now.
10m+20s came in with an average game time of 39:04, with a 90th percentile of 72:30, so notably slower than our 20m+5x30s target, but not exceptionally so. 10m+15s came in faster than 10m+10s with an average of 29:41, so that ones is a little hard to explain other than just a lack of a lot of data for all of these settings. (860 for 10, 431 for 15, and 333 for 20).
Basically my rational for 10+10s is it aligned well with 20m+5x30s in practice. I think thinking of the worst case as “only 10 seconds a move, that’s way too fast” isn’t how to approach it, with Fischer, especially in Go, there’s a lot of opportunity to bank up time. At move 240 for instance, with 10m+10s, each player will have a total of 30m of move time, so the amount of clock time available between the players exceeds the average game by about 2x, so I think in general there’s a healthy amount of time.
That said, we only have a few hundred samples of each of these settings, but the general pattern of “line up your increment to be close to the average move time” seemed to fit well for both blitz and rapid (where we have notably more samples), so I think 10m+10s is a pretty decent choice to pair with 20m+5x30s.
4 Likes
I can’t speak to how IRL tournament clocks might work and why, but every online Fischer implementation I’ve seen, and the description of how Fischer clocks should work, is how we do it where you have your initial time and you add time after each move.
Yeah that description does not describe Fischer as I’ve ever seen it (though I’ve never participated in an IRL chess tournament), I suspect maybe they had some IRL limitations they had to work with and that’s what they came up with.
2 Likes
Thanks, I included it in the table. It’s a bit closer to 5m+5s than 5m+5x30s is.
I think it would be a better and more balanced fit, but if 5m+5x30s is much more popular than 10m+5x20s, I guess it makes sense to still favour 5m+5x30s. After all, the difference seems to be quite minor in practice.
Some people have expressed concerns about players banking Fischer time to the point that they can let their opponent wait for an excessively long time for one particular move in the game, hoping that their opponent ragequits or abandons the game out of boredom.
But the tentative Fischer increments are not very long. A player who blitzes every move in a 10m+10s game might be able to bank some 15 minutes by the later stages of the game for this purpose, but when a player blitzes in a 20m+5x30 game they could do pretty much the same thing (letting their opponent wait for 15 minutes for one move in the game) right from the start of the game.
So I don’t really think this potential issue is unique to Fischer time or that it requires some specific attention, but I suppose some people would be comforted if this potential “exploit” were mitigated by setting the Fischer limit to something like 20 minutes for a full game, 15 minutes for a rapid game and perhaps 5 minutes for a blitz game. Though I don’t know how easily people might run into those limits by accident.
3 Likes
I suppose it depends on the balance between basic time and Fischer increment.
If you use a long basic time with a short Fischer increment, like 60m+10s, then you have about 90 minutes thinking time for the whole game. I suppose you should pace yourself similar to playing a game with 60+1x20s, a byoyomi time setting that has been common for decades in OTB weekend tournaments in Europe. You’d basically try to avoid using up your basic time of 60 minutes before entering the later stages of the endgame.
But you could also use a short basic time with a long Fischer increment, like 10m+30s (which should also give you about 90 minutes thinking time for the whole game). In this case you shouldn’t see it anymore as similar to basic time with some overtime to finish the game (like you would with normal Japanese byoyomi or Canadian overtime OTB tournament games).
At this extreme, Fischer time is indeed a very different kind of time control. Your clock time is more like a water reservoir where you try to keep the water level more or less stable during the whole game. When you think much longer than the increment for a few moves, the water level slowly drops, and when it gets dangerously low, you can play a few quick moves (perhaps a few ko threats) to raise the water level back up again to a safer level. This pacing stays more or less the same for the whole game. There isn’t really a clear transition into overtime.
4 Likes
No, is so different. With byoyomi, byoyomi starts once main time is consumed, and each byoyomi resets if you play inside the byoyomi. In fischer, you have a main time, and each move you do, increase that main time, thats all. If you play inside the increment time, your main time Will increase too. What you mark in the screenshot IS not used in chess online, only tournaments, and is similar to canadian byoyomi.
3 Likes
Yeah, these sound reasonable for automatch settings.
1 Like
I don’t know where this quote comes from, but it is so confused that I wouldn’t rely on it for anything.
Just look at this bit:
if the time delay was five seconds and the time remaining on the player’s clock is ten minutes, they will have ten minutes and five seconds added to their initial remaining time
So by my count that makes the new total twenty minutes and five seconds ??
As I said, bad source for any decision.
1 Like
@anoek :
It seems like the general desire is high enough so that you don’t bump into it in practice, so maybe something like 10m for blitz, 30m for rapid, 120m for full or something?
My guess would be that the cap has a major influence on the duration and speed of the game. So not considering the cap setting when collecting the statistics, or choosing a setting based on the statistics but applying a different cap on top of it, seems a shaky way to make a decision. Do you disagree?
Of course this whole process is fuzzy so maybe it doesn’t really matter.
FWIW, I’m in favor of a meaningful cap like between 100% and 200% of the initial time.
–
Ian
1 Like
Could you, please, show me a better source(s) that tells about how Fischer is used in combination with other time control systems, if at all?
It’s easy to find articles about how increment works, but as a non chess player it’s hard to find a reliable source about how it’s being used. For example, what’s the common time increment in seconds, what are common main times used with Fischer and if capped Fischer perceived as a bad thing by chess players.
Here are some very popular chess Fischer timing courtesy of https://lichess.org/
Numbers left of the + are minutes of starting time, and to the right are seconds added per move. There is no cap.
(Note: due to the relative simplicity and smaller move pool of chess, these specific numbers mostly don’t work for Go, or at least feel very different)
2 Likes
What do chess players have to do with figuring out Go games settings?