Metaverse II

from GalCiv2 Forums
Hmm, some thoughts for the metaverse for GC2.

1) How about an option for us to submit games that won't be counted for score. This will encourage people to submit losses, which you guys can then analyze for better AI. Of course a game submitted this way cannot be resubmitted again at a later date for score.

2) No zeroing of scores. I remember having this argument, and the person I was arguing against tried to argue that having a score in an old block brought to zero is somehow better than having 5% of the score counted, which didn't make sense. Personally I think it would help the longevity of the game if players did not have to feel as if their old game scores would just go to waste.
ideas
5,463 views 3 replies
Reply #1 from GalCiv2 Forums Top

1.) Sounds like a good idea.  I'll add it to our growing list of ideas for the new Metaverse 

2.) IIRC the scores in GC2 only got zeroed when the game went gold, so any scored submitted while overwhelming game imbalances were present in the beta (and beleive me, imbalanced beta's will happen again).  I think we may be able to counter this by forcing the players to only sumbit games that don't count (as in your 1st suggestion) so that we can test the metaverse without worrying about people "running with the cheese" and pumping up their scores with beta imbalances.

That's jsut one idea...the Metaverse is still a few months off, so any comments or ideas before then will be most appreciated

Reply #2 from Stardock Forums Top
Mmm, I meant how games that were too old had their scores drop down gradually, until eventually it became 0% of their original worth. So if I played a game of GC now and submitted it, 2 or so years from now (I forgot the cutoff), if I submit another game, the game I had submitted previously would be worth nothing.
Siege of Centauri
Reply #3 from Stardock Forums Top
More ideas on how to score the new Metaverse

In the old formula, during each 30 day block, your block score was sum of scores/square root of number of games played. This had the effect of lowering your score if you had several good games and then had one bad game. This again, has the effect of causing people to possibly not submit their games (which is bad, since games that aren't submitted can't be analyzed!).

A possible solution would be this. The scores in each 30 day block is sorted in a descending order. Your highest score in the block is worth Score/1. Your second highest score is worth Score/2. Your third highest is Score/3 and so on.

For those who play perfect games (always reach the score cap per game), this won't make too much of a difference. But for those who have a bit more variations in their score, this will give a boost to their score, because their overall score can only go *up*, not down.

Some numbers.

Perfect player playing 10 games under the old system and new system.
Old: 10*60000/sqr(10) = 189736
New: 60000/1 + 60000/2 + 60000/3 + ... 60000/10 = 175738

Average player playing 4 varying games (smaller number because I'm lazy).
Old: (30000+20000+60000+10000)/sqr(4) = 60000 (which means 3 of the games the guy played was basically useless)
New: 60000/1 + 30000/2 + 20000/3 + 10000/4 = 84166

In other words, players are never punished for submitting a game, even a game that's worth 0 points, because it'll just get thrown to the bottom of the stack. However, there is definitely a diminishing margin of utility the more games they play. In fact, the score rise is a bit slower under the new system than the old system if there's little variation.

The only con of this system is that it requires slightly more processing power than the old system (you have to sort the player's scores first), but it shouldn't be too big a problem, because the number of games the average player plays in a 30 day block will usually generate small lists (even one game a day will create a list of only 30 games to sort).

---------------------------------

This system could also be used for the devaluation of old scores. Your current month scores would be worth score/1, the score from the previous month is worth score/2, the score from the month before that is 1/3, etc. Unlike for the 30 day blocks, no sorting's required here. The effect of this system on the score is that it places a much heavier emphasis on your active (current block) scores. However, the devaulation of score slows down drastically as time goes on (change between T (current block) block to T-1 block is 1/2. Change from T-1 to T-2 block is 1/6. Change from T-2 to T-3 is 1/12, etc. But even scores from 3 years ago will be worth a bit (let's say 100000 points in that block, /36 = 2777). In the old system, I think it eventually goes down to 0%.