|Re: Backed into corner on HP48GX|
Message #14 Posted by Don Jennings on 1 Sept 2010, 9:46 p.m.,
in response to message #8 by Gerson W. Barbosa
Yes, if you use most any equal rating you will get 0.5 as the answer, 1200 V 1200, 2400 V 2400, etc. That just means if you are both say, 1600, and you play 100 games, you are pretty sure of winning 50 games and losing 50 games or drawing all games, like most of the games of the Kasparov/Karpov matches, titanic struggles that ended up one game up after something like 130 games, not in one match of course, but they tied a lot of games!
If you solve 'P=1/(10^((O-Y)/400)+1)' for R, assuming you have assigned 1500 for O and Y, you will get P=0.5. Is this what you are looking for? By the way, exponentiation is accessed through the y^x key, but if you prefer you can use the ALOG function instead of 10^x, of course.
So to use Y^x you would go say, 10, enter, Y^x, ((bla bla bla)) for the exponent? or in the equation solver, 10, Y^x, ((Bla))? Does it recognize that last as the proper format for exponents?
I was just sneaking up on the full rating formula, it's a bit more complex than just the win expectancy part, but I wanted to get that right before I finished the rest. The win expectancy is the key, if you don't have that right you are screwed trying to get the change in rating after a game.
Well it sure isn't working like I thought it should, I simplified the expression just to see it work, I entered 4/6 in the equation line of the 'solve' function and hit EXPR= and the exact same thing happened, the hourglass came on for about a half second, goes off and nothing else happens, no answer pops up. What am I doing wrong here?
Ok, figured that one out. You have to hit NXT and do CALC and EVAL, THEN the frigging answer comes out.
Edited: 1 Sept 2010, 10:37 p.m. after one or more responses were posted