>The argument for longer time controls is that it encourages the
>development of new algorithms. New algorithms are usually slower. It
>might take 10 man hours to quickly code up a new idea. Sure we can
>optimize it to run 10 times quicker, but that takes another 90 man
>hours. We want to see how well an idea works before spending all that
>effort.
One thing I would *really* like to know is how to make something
10 times quicker using only 90 man hours. :-)
My experience is along the following lines:
- Faster games give you faster feedback about your ideas.
- There are limits to the value of even faster time controls.
- The purpose of new algorithms is to make better use of time,
so value should be revealed under fast time controls.
This is a matter of opinion or belief, so there is no right and wrong here.
I am just saying that I value rapid experimentation and a good balance
between precision and volume of feedback. For me, the 5-minute controls
work well.
>If the same bot plays both 5 minute games and 20 minute games, and the
>server can show different ratings for each "venue", it would show how
>well that program's algorithms scale.
Now this I understand. I have a very inadequate computer, and I could
spare myself the expense of buying a faster computer by playing slower
games.
But on balance, that is a nice-to-have feature that rarely matters to me.
If I want to test scalability, I will scale *down* and assume that scaling
up extrapolates reasonably.
For the reasons given above, Pebbles will play in the 5-minute venue.
This doesn't mean that other venues are bad. They might be useful.
But I would like to have a large number of opponents in my
preferred venue.
_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/
