Response to leaderboard design question

You're right. I'm tired. I shouldn't be sat around breaking MARP at this time in the morning...

My picks for the competition I'm not sure about. Should I aim to pick the games I'm most likely to beat you guys at, or games I like to play? Hmm. Either way qbert's qubes has to be in there - you never did manage to reclaim that one, did you?

I've just been playing with the leaderboard scoring mechanism a little more. I coded up the idea I mentioned a few days ago, where the player's percentage of the highscore is multiplied by 0.85 to the power (position - 1). Check it out & tell me what you think. It's swapped German Krol and JoustGod in positions 2 and 3 of the table, but I can't work out why that is. JoustGod has more recordings, but German has more 1st places. I'm wondering if JoustGod has enough recordings which score 15% or less than the highscore, which would have counted for zero previously, and now count for something more than zero...

Whatever - I've tidied the script up a little, so it's relatively easy to slot new scoring code into place...

Notice that the search pages now report the number of leaderboard points you've earned, rather than the flat percentage (and they assume that the bonus is 15%).

Also, the search pages can now sort on leaderboard points awarded, as well as score, date, etc.

It's past my bed time...

Chris.

[ Previous | Next ]