Average play time?

Originally posted by erice1
See thats the point. It doesn't matter how 'many moves' you make. If your openent takes 21 days to move but you respond in 15 minutes. Your move time is 15 minutes if you only make the one move in the month.

however, your openents time will be 21 days if thats the only move he makes in the month.

I think its very valuable info.

Sorry I don't get it.

Could you give a worked example with multiple games and multiple moves per game per month?

D

I think he wants to see the average time between your opponent's move and your move in all of your games.

Problem is, I'm not sure that's a very useful number. I could be moving every 5 minutes in dozens of 1/0 games, and then take on a 21-day game and wait the full 21 days every single move.

In that case you are going to be annoyed because the statistic didn't correctly predict what you expected it to.

Really, you should just expect that everyone is going to use the full length of the time control, and then choose the time control accordingly. Any other method you try to use is assuming some imaginary constraint on your opponent that doesn't really exist.

Just to try this out with some actual numbers:

Say I make 100 moves in a particular month. 99 of them I make in 5 minutes and 1 of them I take 21 days.

So now my average time is (99*5 + 1*21*24*60)/100 == 307 minutes or just over 5 hours.

I'm not sure this number is really useful, because it's very inaccurate both for my fast games (the vast majority) and for my one rare slow game.

Originally posted by incandenza
Just to try this out with some actual numbers:

Say I make 100 moves in a particular month. 99 of them I make in 5 minutes and 1 of them I take 21 days.

So now my average time is (99*5 + 1*21*24*60)/100 == 307 minutes or just over 5 hours.

I'm not sure this number is really useful, because it's very inaccurate both for my fast games (the vast majority) and for my one rare slow game.

so thats exactly what I'm looking for.

There are 'corner cases' where someone may sit on a game for 21 days to avoid a checkmate, but most people play consistently.

Originally posted by erice1
so thats exactly what I'm looking for.

There are 'corner cases' where someone may sit on a game for 21 days to avoid a checkmate, but most people play consistently.

But this is not just corner cases, it's the normal case. Normally a player will manage 1-day games very differently than 21-day games. By mixing them together in the average you're going to get a number that is way off for both of them and essentially meaningless in terms of predicting anything.

But basically I find the concept ill-founded because it's based on the idea that you can expect people to move faster than the time controls. That expectation is only going to lead to frustration and complaints when it's inevitably not met.

Maybe there should be a warning telling you what the total length of a 40-move game will be if both sides use all the available time.

Originally posted by erice1
so thats exactly what I'm looking for.

There are 'corner cases' where someone may sit on a game for 21 days to avoid a checkmate, but most people play consistently.

As incandenza states, these aren't corner cases.
I have 21 day tournament games which I generally move 20 days into my TO.

I have 7 day tournament games in which I move somewhere between 4 and 7 days into my TO.

I have 3 day tournament games which I move after 1 to 3 days.

I have friendly games where I mightn't move for a couple of months.

And then I have won games, where I blitz out the finish.

Your stat achieves nothing, apart from adding massively to the strain on the server. The time controls are the only things that will tell you how often your opponent is going to move.

D

Originally posted by Ragnorak
As incandenza states, these aren't corner cases.
I have 21 day tournament games which I generally move 20 days into my TO.

I have 7 day tournament games in which I move somewhere between 4 and 7 days into my TO.

I have 3 day tournament games which I move after 1 to 3 days.

I have friendly games where I mightn't move for a couple of months.

And ...[text shortened]... controls are the only things that will tell you how often your opponent is going to move.

D

ok. I guess I was working under the assumption that people enjoyed playing chess so they sign into the system, see that they 'can' move and do. Personally, I really enjoy seeing that its 'my move' and seldom sign out with a game still pending.

As to overhead, its minimal. The system is already logging that I moved (a counter on how many moves I've made this month). To log the time between my move and my opponents last move is one additional line of code and a new field in the db.

Sounds like its not going to happen.

Originally posted by erice1
ok. I guess I was working under the assumption that people enjoyed playing chess so they sign into the system, see that they 'can' move and do. Personally, I really enjoy seeing that its 'my move' and seldom sign out with a game still pending.

As to overhead, its minimal. The system is already logging that I moved (a counter on how many moves I've made ...[text shortened]... additional line of code and a new field in the db.

Sounds like its not going to happen.

What makes you think you like chess more than the people who realize your number means nothing?

P-

Originally posted by erice1
ok. I guess I was working under the assumption that people enjoyed playing chess so they sign into the system, see that they 'can' move and do. Personally, I really enjoy seeing that its 'my move' and seldom sign out with a game still pending.

As to overhead, its minimal. The system is already logging that I moved (a counter on how many moves I've made ...[text shortened]... additional line of code and a new field in the db.

Sounds like its not going to happen.

You have 6 games in progress. Tell the people with 1000+ games going that you enjoy the game more than them because you move in all your games at one sitting. 🙄

You underestimate the overhead. Its one additional calculation, to be done everytime anybody on the site makes a move. Its an additional read from the database and right to the database everytime a move is made. That could be over 100,000 (could be many more, I don't know) added calculations per day.

For a meaningless number, that's a lot of overhead.

D

I have developed a list of players who LOVE chess. In order, top to bottom:

http://www.timeforchess.com/core/playertable.php?state=asc&buttonno=2

Originally posted by Phlabibit
I have developed a list of players who LOVE chess. In order, top to bottom:

http://www.timeforchess.com/core/playertable.php?state=asc&buttonno=2

Great list.. thanks for sharing.

It seems nobody likes my idea and its turning a bit hostile.

This will be my last post and I'll just go play my games.

Originally posted by erice1
Great list.. thanks for sharing.

It seems nobody likes my idea and its turning a bit hostile.

This will be my last post and I'll just go play my games.

Don't agree = hostile.

I just found it a little condescending that you equate moving in your 6 games as "I guess I was working under the assumption that people enjoyed playing chess so they sign into the system, see that they 'can' move and do."

That's a silly thing to say in my mind, and I said as much.

That list I posted should help you a lot, I used to play 2 games with the MAP leader almost every month... but don't really have that time anymore. PM them, they are most often always up for a fast game or 2.

Give it a try, and sorry if you feel my posts were hostile... I wasn't intending to be mean, just disagreeing with what you said.

P-

Originally posted by erice1
Great list.. thanks for sharing.

It seems nobody likes my idea and its turning a bit hostile.

This will be my last post and I'll just go play my games.

I would suggest playing other non-sub scum at the minimum time settings, they only have a few games going and in my experience seem to move quicker.

Problem solved.
You're welcome. 🙂


As a bonus to this approach you can't be hornswoggled by the vacation system either. 😉

Originally posted by Dr Strangelove
As a bonus to this approach you can't be hornswoggled by the vacation system either. 😉

Hornswoggled... mmm... I like that word!

Otherwise, I think "at the minimum time settings" is the main clue. If someone offers a 21/28 game, it's not likely that xe plans to move fast in that game, no matter whether xe's a sub or non-sub and what xyr average move time is.

Percent of timeout used between moves

Average =(total timeout - timeout remaining)/(total timeout)

Example
Average = (7-3.5)/7 = 50%
Average = (3-1.5)/3 = 50%

On average, Player X uses 50% percent of his timeout before moving