Author: Robert Hyatt
Date: 09:31:56 06/13/01
Go up one level in this thread
On June 13, 2001 at 10:50:30, José Carlos wrote: >On June 13, 2001 at 09:45:09, Robert Hyatt wrote: > >>On June 13, 2001 at 06:14:37, José Carlos wrote: >> >>> >>> The interesting thing of the graph is the shape of the curves. Although the >>>x-axis scale is not constant (which makes the "Crafty gets more positions almost >>>linear" statement not correct) the shape of the curves show different strength >>>increase with time for the three programs. >>> Of course, you can argue that this is just a test, and doesn't prove anything >>>itself. And I agree with that. But it will mean something _if_ further tests >>>give similar results. >>> >>> José C. >> >> >>Note that linear does not mean "perfect". IE if you search twice as long >>and every time you double the time, you double the number you get right, >>that is linear. But if you double the search time and you get 1.2 times >>more correct answers, _that_ is also linear. Linear is a straight line. It >>doesn't have to have a slope of 45 degrees.. > > I know what linear means, Bob. If you look at the bottom line of the graph: > > > ---------------------------------------------------------------- > 5s 20s 1m 3m 10m > > you can see that the difference (in time) between equidistant (is this word >correct in english?) points is not constant: > 20-5 = 15 sec > 60-20 = 40 sec > 180-60 = 120 sec > 600-180 = 420 sec > > So, I see here a logarithmic growing. Am I wrong? > > José C. Nope... not wrong at all. but I haven't tried to plot the data either so I don't know if the curve is a straight line or an exponential that is flattening out as time goes up. My point was that the linear relationship most people talk about is a line with a slope of 45 degrees. 2x faster == 2x more correct answers, for example. 2x faster -> 1.1X more answers is also a perfectly straight line... > >>IE in the "Crafty goes deep" experiment done by Monty Newborn, the number was >>something like 15% better for each additional ply of depth. That is still a >>linear curve.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.