Originally posted by greenpawn34That's what I thought... You can't do that, Geoff. It has to be a weighted average. 🙂
I added all the % wins by White and divided by 20 then did the same for Black
then did the same for the draws.
20 is the number of different first moves that have been played in that DB.
(Lies, damn lies, and statistics)
Originally posted by greenpawn34you added it as if 1.a3 was played just as many times as 1.e4.
Seems a logical way to do it.
I sat and done in my head (cept the diving by 20 bit then I used a calcualor)
I was quite pleased when everything equaled to 100.
percentages always add up to 100% by definition. otherwise they're not 1/100ths.
Originally posted by wormwoodSo what if 1.a3 is played less than 1.e4.
you added it as if 1.a3 was played just as many times as 1.e4.
percentages always add up to 100% by definition. otherwise they're not 1/100ths.
If you are saying weak players tilt the numbers because of their odd openings
then remember that they two have Black and will tilt them back.
You cannot apply your weighted thing (what ever that means) to this equation.
These are people you are talking about, human beings not apples.
I know a total % should = 100 it was how I know my running maths were correct.
I'll give Ruxton a call around about noon when he usually gets up he's one
of these doctors at maths. He knows about these things.
Originally posted by greenpawn34Use the formula from Gatecrasher's thread, it's the accurate one.
So what if 1.a3 is played less than 1.e4.
If you are saying weak players tilt the numbers because of their odd openings
then remember that they two have Black and will tilt them back.
You cannot apply your weighted thing (what ever that means) to this equation.
These are people you are talking about, human beings not apples.
I know a total % s ...[text shortened]... noon when he usually gets up he's one
of these doctors at maths. He knows about these things.
Originally posted by greenpawn34If we use your non-weighted method on the RHP master database, for the 16 first moves we get:
Seems a logical way to do it.
I sat and done in my head (cept the diving by 20 bit then I used a calcualor)
I was quite pleased when everything equaled to 100.
White = 28.6%
Black = 36.9%
Draws = 34.4%
So the RHP masters must blunder more than the under 1900s!. 🙂
Originally posted by greenpawn34No, we're actually talking about games. It being in percentages instead of games might be throwing you off. If the database had something like this
You cannot apply your weighted thing (what ever that means) to this equation.
These are people you are talking about, human beings not apples.[/b]
e4 -
white - 160,130 wins
draws - 34,811
black - 153,168 wins
Na3 -
white - 19 wins
draws - 4
black - 26 wins
Would that make more sense? It would be absurd to take the percentages of e4 and Na3 treated them equally. You have to add up all the wins for for each color and the draws and then divide by the total games.
Hi guys.
Of course you are both wrong.
Just taking the first 16 of a DB is pure folly and you are manipulating the figures.
Look at it this way.
If had a DB of all the games played on every site on the net and ask it
to find only mates in two moeves then it would come back with:
100% Black and 0% White. So you cannot chop the moves. That's cheating.
---------
All 20 legal moves are displayed with the % of wins, losses and draws.
To start favouring 1.e4 over 1.Na3 just because players have chosen that
as the first move the most times is misleading.
Following that path you will have to see all the 1...e5 replies and all the 1..e6
games and before you know where you are you have over I million
fraction of a % to work with.
Originally posted by SmittyTimeAll you have to do is look at your own stats. Your winning percentage is 63% as white and 54.4% as black.
For some reason I am convinced I am better playing black. I'd like to think it is becuase I am better at reacting than initiating, but its probably becuase I don't know any openings.
Originally posted by greenpawn34GP, I think you misunderstood me. I wasn't saying that I was chopping the games off to the first 16 moves of the game. I was saying that in the master database, there were 16 different first moves by White.
Hi guys.
Of course you are both wrong.
Just taking the first 16 of a DB is pure folly and you are manipulating the figures.
Look at it this way.
If had a DB of all the games played on every site on the net and ask it
to find only mates in two moeves then it would come back with:
100% Black and 0% White. So you cannot chop the moves. That's cheating.
Oh I see - you said 16 moves.
So yes according to that selective DB of master games Black does better.
I can show you other master DB that have White doing better.
However in this case we are talking about a closed DB here on good old RHP
so comparing the 1400 DB to a DB of master games is not really fair.