A useful stat on a players profile would be:
Average number of moves per game per day and/or
Average number of days to complete each game.
It may allow players to match up with opponents who play at roughly the same pace.
I'm sure more accurate numbers can be derived from actual game data (if start and end dates are kept) but here are the formulas required for an unbiased suck-thumb (with many assumptions) for anyone interested in assessing the speed of opponents prior to playing them:
Firstly we are given
d = Days as a member (ie today-joined+1)
c = Completed games
a = Max(1, In Progress games)
t = Max(1,Total moves)
Thus m, the average moves per game is
m = t / (c + a/2)
From this we get x, the average number of moves per game per day
x = t /da + m/2d
And y, the average number of days to complete each game
y = m/x
By way of comparison:
Edmond Dantes (highest rating) : x = 0.66 y = 34
Feivel (vocal rival of ED): x = 0.40 y = 71
Rhymaster (most active): x = 1.81 y = 17.4
Gatecrasher (obsessed newbie) : x =5.65 y=5.1
Have fun.
all this sounds very good, but... don't forget about 2nd moments (variance):
I think the move frequency is highly correlated with the T/O setting. so perhaps I play two games with 21 T/O and I move twice a month there and in all the other 3 day T/O games I move every day.
another "but": the move frequency is highly dependent on my opponents movingspeed: Imagine I move always within 5 min, my opponet in 6.9 days.
Nevertheless your estimators are very interesting. perhaps just all MY thinking time should be used for calculations. But this means a lot more work for our sweet server and some coding for our admins 😉 😀
interesting discussion...
th
Moves per day sounds like a pretty good and relatively simple stat to add to a user's profile page. For example, I signed up just short of 4 weeks ago (so 26 days). I have made 562 moves at the time of this post. Both these figures are easy to find. I have thus made an average of almost 22 moves per day. Since I have tended to be playing 8-10 games at a time, that is roughly 2 moves per day per game I have been playing (which is another stat that you might want to consider). (That is, I just mean to reinforce Gatecrasher's suggestions, not to say that this is different. It's just slightly more compicated to do moves per day per game, or maybe it's not. Depends on what figures are how easy to pull out of the database.)
On the other hand, one of my games has been going on for more than 3 weeks now since that player (provisional) has reduced his play frequency in our game to one move every 4-6 days after losing his Queen. 🙁 Not that it really matters all that much, but a game like that could throw a monkey wrench into someone's stats.
Paul
as Russ just said today:
little effect, big traffic costs... 🙁
and another thing about moving speed (and why measuring it is not so simple): I tend to move fast in the opening (and the end game).
And when I am behind a bit (weaker position) I tend to move slower as well...
same with you?
th
Some good points have been made, which is why a speed stat derived from known information, like the one I described, would be a thumb-suck. But over a reasonable period, and given an average number of outstanding matches, the stat is far better than nothing.
If there was a way of quantifying the average period that a player takes to make a move after his opponent has moved we would really be on to something.
Originally posted by thireI like the idea of having more stats as well.
as Russ just said today:
little effect, big traffic costs... 🙁
and another thing about moving speed (and why measuring it is not so simple): I tend to move fast in the opening (and the end game).
And when I am behind a bit (weaker position) I tend to move slower as well...
same with you?
th
As said stats give you a just _rough_ idea but better than nothing.
Concerning traffic costs.
I believe that there is not much more traffic just a few more metrics in the profile page. On the one hand calculating the real-time stats may be ressource intense for the database but on the other hand that kind of stats don't have to be up to date every second. If they are calculated just once a day or even once a week they would still be fair enough.
my 2 cents,
Gotti