New Time System Hourglass

Kaya.gs · #1

Hey Guys. 0.20 release included Hourglass. A couple of games were played with it and im actually very surprised at how good it is.

Quote:

Hourglass

In Hourglass each second a player takes is added to his opponent time. A player loses after the difference between his clock and his opponents exceeds the differential time. Kaya.gs has a slight modification: if you play, you will have at least have 10 seconds to think your next move.

Two games were played with this system with 10 minutes main time and one of them lasted an hour. It seems to be an excellent time system for the players.

Himiko here played one of them, maybe she can give an opinion.

I had thought of making this time system not-rateable because it seems too "funny" but in practice it has been a very serious option so far.

The main advantage of hourglass is that neither player takes excessively more time than the other one. Which means you dont get the feeling that "your opponent is slow".

The disadvantage i see is mainly for observers. A game with 10 minutes Hourglass is potentially a 20 minute period byoyomi game! So it can be really really long.

Any thoughts?

hyperpape · #2

Definitely a cool idea. It might be too complicated to catch on, and people might stick with the defaults, but I still like it. I'm wondering if it might be awkward, since if you choose a short differential, then your opponent can force a very short game by playing fast, while if you choose a differential near your desired game length, you're agreeing to a game 50% longer than the one you want. (If you say 30 minutes, your opponent can play for 60 minutes).

topazg · #3

hyperpape wrote:

Definitely a cool idea. It might be too complicated to catch on, and people might stick with the defaults, but I still like it. I'm wondering if it might be awkward, since if you choose a short differential, then your opponent can force a very short game by playing fast, while if you choose a differential near your desired game length, you're agreeing to a game 50% longer than the one you want. (If you say 30 minutes, your opponent can play for 60 minutes).

Only once though. After that, unless both players are doing it, that guy will be on a few seconds sudden death each time

badukJr · #4

So its like double penalty time? You can already think during your opponents time.

SpongeBob · #5

Very cool ... will have to try it!

However, as I am usually the slower of the two players, it is probably not the ideal time system for me (but more for my opponents ;-)

).

Bill Spight · #6

Hmmm. How about if the time runs down, until you reach byo-yomi?

How about something like this? Each player starts with 30 min., total 1 hr. Plan for 240 moves in that hour, so each move averages 15 sec. Subtract 15 sec. from the time to be added for each move, but never subtract time. So if the first move takes 1 min., add 45 sec. to the opponent's time. If it takes 5 sec., do not add anything. (With this scheme it will probably take considerably longer than 1 hr. for the total time to run down.)

As for byo-yomi, say that byo-yomi is 10 sec. per move. If a player's time after the opponent's play is less than 10 sec., make it 10 sec.

One feature is the possibility of long byo-yomi in difficult situations. If one player takes 45 sec. for a move, his opponent gets at least 30 sec. to reply.

hyperpape · #7

topazg wrote:

Only once though. After that, unless both players are doing it, that guy will be on a few seconds sudden death each time

What does this mean? I play for thirty minutes, my opponent plays for sixty. That's the game. I don't mean he can take thirty minutes for one move, but just thirty minutes longer over the course of the game.

My point isn't that my opponent can gain an advantage, but just that I'm either forced to accept the possibility of a quite short game or a quite long one. My opponent can always play all his moves in five minutes, limiting me to the main time plus five.

So it's only ok for when you really do want to be quite flexible about how long a game lasts. I guess that's just obvious, but it does seem like a downside.

Tami · #8

I`m Himiko on Kaya (and KGS and WBaduk), and I played one of the first hourglass games, my opponent being Practise.

It felt a lot like Fischer - you can get the novel sensation of seeing time accumulating on your clock, and the game reached an assymetrical stage where I had a lot of time, while Practise was down to his last minute. However, unlike absolute time or even Canadian time, it does not appear likely that you can win by simply by moving quickly - all the opponent has to do is play one good or surprising move and you lose the benefit of any pressure you may have created because you will need to think.

The key will be finding a time allowance that makes for a suitably paced game. Our game was began with ten minutes each (therefore there were 20 minutes in the hourglass), and lasted for about an hour. I played very quickly, though, and had I taken my time, the game might have lasted two or three hours.

My suggestions:

Blitz: 30 seconds each (1 minute hourglass)
Quickplay: 90 seconds each (3 minute hourglass)
Short: 3 minutes each (6 minute hourglass)
Long: 5 to 10 minutes each (10 to 20 minute hourglass)
Serious: more than 10 minutes each

Time management is a crucial skill, and the game I played helped me to appreciate the benefit of having practiced with a lot of life and death problems. For sure, I made a lot of reading mistakes, but I was able to read sufficiently well to play comfortably at the time set and to use its properties to my advantage.

So, all in all, hourglass is defintely not a "Mickey Mouse" time control. It's fun, it's complex, and it's only on Kaya.

ez4u · #9

I have played a number of hourglass games IRL, using a chronos game clock. I think the max time we ever used was 90 seconds each. Since you are tracking and restricting the difference in time used between the two players, it is not immediately obvious to me why long-period hourglass would provide a more enjoyable gaming experience than more common timing systems.

The 10-second delay is a good idea since it is quite common for one player to catch the other in a time squeeze without it. Hopefully the delay would be a modifiable option, however, rather than a fixed attribute.

topazg · **#10**

hyperpape wrote:

topazg wrote:

Only once though. After that, unless both players are doing it, that guy will be on a few seconds sudden death each time

What does this mean? I play for thirty minutes, my opponent plays for sixty. That's the game. I don't mean he can take thirty minutes for one move, but just thirty minutes longer over the course of the game.

My point isn't that my opponent can gain an advantage, but just that I'm either forced to accept the possibility of a quite short game or a quite long one. My opponent can always play all his moves in five minutes, limiting me to the main time plus five.

So it's only ok for when you really do want to be quite flexible about how long a game lasts. I guess that's just obvious, but it does seem like a downside.

This doesn't seem possible unless you have an hourglass that's 30 minutes long (which is _huge_ for hourglass timing) - the time system forces your total time to be no greater than your opponents time + the max in the hourglass. In a more typical setting of maybe a 3 to 5 minute hourglass the max time he could have taken is 33 to 35 minutes if you took 30.

Bill Spight · **#11**

Maybe I am misunderstanding, but it sounds like if each player starts with 6 min., then if my time drops below 3 min., I lose, because that means that my opponent has more than 9 min., and the differential is greater than 6 min. That's equivalent to a system where we do not care about the differential, we just want our time to stay above 0, and each player gets 3 min. Since that is simpler and equivalent, that is how I think that it really works, no?

Kaya.gs · **#12**

Bill Spight wrote:

Maybe I am misunderstanding, but it sounds like if each player starts with 6 min., then if my time drops below 3 min., I lose, because that means that my opponent has more than 9 min., and the differential is greater than 6 min. That's equivalent to a system where we do not care about the differential, we just want our time to stay above 0, and each player gets 3 min. Since that is simpler and equivalent, that is how I think that it really works, no?

No, you lose when your clock drops to 0. Your click increases its time only when your opponent is thinking. So if you have 20 seconds, you can only get it to 30 seconds if your opponent thinks 10 seconds more than you.

ez4u wrote:

I have played a number of hourglass games IRL, using a chronos game clock. I think the max time we ever used was 90 seconds each. Since you are tracking and restricting the difference in time used between the two players, it is not immediately obvious to me why long-period hourglass would provide a more enjoyable gaming experience than more common timing systems.

The 10-second delay is a good idea since it is quite common for one player to catch the other in a time squeeze without it. Hopefully the delay would be a modifiable option, however, rather than a fixed attribute.

I have absolutely no experience with hourglass. I didnt even play one myself. I got the feeling that 5 minutes hourglass (what i thought was minium) is a bit too much, however, as the game progresses those 5 minutes easily vanish.
On a non-blitz game, spending 3 or 4 minutes in a move is can be very valuable. 90 seconds hourglass would make that impossible.

Of course, usage and experience will give us the better numbers, as we know now that 10 sec byo yomi is blitz and 30 sec is standard, there is no standard yet for fischer or hourglass.

hyperpape wrote:

Definitely a cool idea. It might be too complicated to catch on, and people might stick with the defaults, but I still like it. I'm wondering if it might be awkward, since if you choose a short differential, then your opponent can force a very short game by playing fast, while if you choose a differential near your desired game length, you're agreeing to a game 50% longer than the one you want. (If you say 30 minutes, your opponent can play for 60 minutes).

Indeed the most outstanding property of this time system is that is the one that can give the longest possible games (out of the known time systems). As if both players used maximum time, even a 5 minute hourglass could last almost 10 minutes per move.
Of course that is worst case scenario and ridiculous, but it certainly gives room for a really long game.

That was my initial observation in the original post. However for the players its a really good thing because it means that they can play a really long game without the concern of boring your opponent. Its never abusive because you are taking almost as much time as your opponent no matter what.

Archivist · **#13**

Having watched the first hourglass game on Kaya, I got the impression that shorter times would make for a more interesting game. And that it is a worthwhile, functional time system as it is. We didn't see what happens to the overall time when a player gets down to the last 10 seconds, but on the whole it seems to reward reading quickly and playing moves that cause the opponent to think.

Bill Spight · **#14**

Kaya.gs wrote:

Bill Spight wrote:

Maybe I am misunderstanding, but it sounds like if each player starts with 6 min., then if my time drops below 3 min., I lose, because that means that my opponent has more than 9 min., and the differential is greater than 6 min. That's equivalent to a system where we do not care about the differential, we just want our time to stay above 0, and each player gets 3 min. Since that is simpler and equivalent, that is how I think that it really works, no?

No, you lose when your clock drops to 0. Your click increases its time only when your opponent is thinking. So if you have 20 seconds, you can only get it to 30 seconds if your opponent thinks 10 seconds more than you.

Your explanation seems to conflict with the description given:

Quote:

In Hourglass each second a player takes is added to his opponent time. A player loses after the difference between his clock and his opponents exceeds the differential time.

Notice that there is nothing about losing when your clock drops to zero. You lose when the difference between your clock and your opponent's clock exceeds some value.

Kaya.gs · **#15**

Bill Spight wrote:

Notice that there is nothing about losing when your clock drops to zero. You lose when the difference between your clock and your opponent's clock exceeds some value.

Its the same thing.

Let me give you an example.

Hourglass 10 min

Player A Player B
05:00 05:00
04:50 05:10
05:02 04:58
02:00 08:00
00:10 09:50
00:00 10:00 Player B + Time

badukJr · **#16**

Tami wrote:

My suggestions:

Blitz: 30 seconds each (1 minute hourglass)
Quickplay: 90 seconds each (3 minute hourglass)
Short: 3 minutes each (6 minute hourglass)
Long: 5 to 10 minutes each (10 to 20 minute hourglass)
Serious: more than 10 minutes each

You have a weird definition of 'Short'. If each player uses 3 minutes for each move to move 200, that is 10 hours.

A 10 minute each game can easily last more than a day. This time system is not really suited for tournaments.

Kaya.gs · **#17**

badukJr wrote:

You have a weird definition of 'Short'. If each player uses 3 minutes for each move to move 200, that is 10 hours.

A 10 minute each game can easily last more than a day. This time system is not really suited for tournaments.

Why wouldnt it be suited for tournaments? Its not suited to one-day tournaments, because you cant tell when a game is going to end but whats the issue with any other kind?

Bill Spight · **#18**

Kaya.gs wrote:

Bill Spight wrote:

Notice that there is nothing about losing when your clock drops to zero. You lose when the difference between your clock and your opponent's clock exceeds some value.

Its the same thing.

Let me give you an example.

Hourglass 10 min

Player A Player B
05:00 05:00
04:50 05:10
05:02 04:58
02:00 08:00
00:10 09:50
00:00 10:00 Player B + Time

That's what I said in post #11. So why keep track of the difference in clocks?

Kaya.gs · **#19**

Bill Spight wrote:

Kaya.gs wrote:

Bill Spight wrote:

Notice that there is nothing about losing when your clock drops to zero. You lose when the difference between your clock and your opponent's clock exceeds some value.

Its the same thing.

Let me give you an example.

Hourglass 10 min

Player A Player B
05:00 05:00
04:50 05:10
05:02 04:58
02:00 08:00
00:10 09:50
00:00 10:00 Player B + Time

That's what I said in post #11. So why keep track of the difference in clocks?

Ah , merely to use the same interface as any other clock. We thought of making an actual hourglass animation , but it takes some effort and it has limited value. I guess we will do it some time in the future.

Bill Spight · **#20**

Bill Spight wrote:

Kaya.gs wrote:

Bill Spight wrote:

Notice that there is nothing about losing when your clock drops to zero. You lose when the difference between your clock and your opponent's clock exceeds some value.

Its the same thing.

Let me give you an example.

Hourglass 10 min

Player A Player B
05:00 05:00
04:50 05:10
05:02 04:58
02:00 08:00
00:10 09:50
00:00 10:00 Player B + Time

That's what I said in post #11. So why keep track of the difference in clocks?

Kaya.gs wrote:

Ah , merely to use the same interface as any other clock. We thought of making an actual hourglass animation , but it takes some effort and it has limited value. I guess we will do it some time in the future.

But isn't it simpler just to keep track of 0?

New Time System Hourglass

Who is online