Metaverse scoring weirdness

from Stardock Forums
Something appears to be wrong, either we are no longer
informed on the math behind the metaverse or its very
buggy.
I calculated the top 3 players scores using excel
not counting any decay

Staffa Actual 82479,by math 49559.15
Ray... Actual 74665 by math 42075.42
Ponti. Actual 70495 by math 45902.79

1,488 views 69 replies
Reply #1 from Stardock Forums Top
While trying to figure out what was going on, I had
a somewhat unrelated good idea.

The overall should be the sum of your individually scored
map sizes.

This would reward people who played different map sizes
quite nicely as the sqrt penality would be less for
people who played on all map sizes then someone
who only played on one map size.

This btw is not what is currently going on, I
already checked the math for that.

Reply #2 from Stardock Forums Top
Try grouping scores by month.

Compute the scores of the groups according to the current formula and then sum up the scores for every month. I'll do this later if nobody beats me to it.
WindowBlinds - Free yourself of the default settings, customize your desktop.
Reply #3 from Stardock Forums Top
Tried it, came out to 91k for me. Maybe with decay
this is what is going on.

Btw, insightful idea on what might be going on.

Reply #4 from Stardock Forums Top
Also of interest, your total metaverse score listed on the metaverse rankings page is different that what is shown when you click on your name which happens to be the same score as is shown next to your name in the empire members screen.

I wonder if any of this will even get addressed though with the metaverse budget exhausted so is it worth even discussing?
Reply #5 from Stardock Forums Top
IceMole suggested the follwing sometime back. Someone should verify:

#1213 by Citizen IceMole - 7/9/2003 5:34:53 PM

While I have not verified the exact formula for grandfather scores for myself, I do agree that old scores on the overall list have been compensated for somehow.

Metaverse overall score seems to be calculated as follows:

Scores are grouped into blocks by their age/degradation.

Within each block, the sum of the scores is divided by the square root of number of games in that block.

The results of all the blocks are added together to give total overall score.

So, if my alter ego NiceMole submits the following games:

Date Score Degraded to Current Value

3/29 20 70% 14
4/4 36 70% 25.2
4/26 207 80% 165.6
6/30 8260 100% 8260

Block 70%: (14 + 25.2) / 2(sqrt) = 27.7
Block 80%: 165.6 / 1(sqrt) = 165.6
Block 100%: 8260 / 1(sqrt) = 8260
Total: 27.7 + 165.6 + 8260 = 8453.3

This total would be the one given on the player's page. The one shown on the list is often slightly higher, which I think may have something to do with the timing of the application of score degradation
End of quote
Reply #6 from Stardock Forums Top
I have calculated Staffa's score and come up with only 80024 of his actual 80150. Further investigation has shown that my method misses by a small fraction of a percent on most scores with more than a few games.
Reply #7 from Stardock Forums Top
Did you weight the different versions, IceMole?
Reply #8 from Stardock Forums Top
I do not believe there is any weighting based on version at the moment.

As an example, LarryHal with score 6986.4, rank 569:
Date - Version - Score - Degradation - Current Value
4/7....1.02......93......80%...........74.4
6/23...1.05......6912....100%..........6912

Block 80%: 74.4 / 1(sqrt) = 74.4
Block 90%: 0
Block 100%: 6912 / 1(sqrt) = 6912
Total: 74.4 + 0 + 6912 = 6986.4

This corresponds exactly to his player page score, so there does not seem to be any weighting going on.
Reply #9 from Stardock Forums Top
Excellent work, IceMole.

This shows that the Metaverse computations have indeed been changed so that older players are not penalised for the small scores obtained from playing prior versions.

This would explain why we are not seeing the expected scenario of good newbies overtaking active senior players on the Metaverse.
Reply #10 from Stardock Forums Top
Actually Ray, I'm not so sure this is a change. I have been playing with score calculations since I started in the metaverse, and while I don't have the old numbers to back it up, I have noticed that my score at least has been higher than the classic formula since my fifth game. Looking now, I see that this is when my first game first aged. I think the reason people are noticing now is that their inflated 1.05 scores are beginning to affect multiple blocks, working more efficiently.
Reply #11 from Stardock Forums Top
That's possible, IceMole. Many of us have just been assuming that the sqrt applies to all games instead of in groups. The decay just confuses the matter further.
Reply #12 from Stardock Forums Top
I've just been playing with the list scores and having success by applying half the degradation, or 5% of the base per month. Unfortunatly, it still seems to miss by a fraction of a percent when more than a few games are involved.
Reply #13 from Stardock Forums Top
Holy burning effigies of Barny!

How did I get to be #5 for average score? The math around here... wait a minute... THERE IS NO MATH!!! It's all random, isn't it? Now it all makes sense. I am one with the Meta. Oooooom... Oooooom... repeat after me... Oooooom... Oooooom... there is no math... Oooooom... Oooooom...
End of quote


Can I have just a little laugh... pleeeeeease? ;)
Reply #14 from Stardock Forums Top
Oh, I have been laughing maniacally for some time now. I think I left my sanity behind after the first couple thousand calculations. LOL
Reply #15 from Stardock Forums Top
Curious, that does mean old scores can not hurt you and
would explain why I played 1 game for 19000 points and
my overall went up 14000.

I havnt played many games lately, my last game was just
the 4th game in this block.

Ray is on about his 15th game in this block and
consquently getting far less overall point increase
for his work. Hence gaining 2000 points for his 15k
games.
Reply #16 from Stardock Forums Top
This actually means the metaverse seriously favors
old timers as it will be very hard for new players to
catch up.

Odd how a little detail in the calculations can so
dramatically change the outcome, wonder why they
never bothered to let us know this. It eliminates
quite a few major concerns people had with the metaverse
while introducing a couple others(which oddly favor
me this time, rather then hurt me, so I will let
new players do the complaining this time)

Reply #17 from Stardock Forums Top
Also means that at the start of each new month,
overall scores will jump up very quickly and of
course scores during the end of months will tend to
not do much to the metaverse.

Reply #18 from Stardock Forums Top
Actually what is really means is that you should only submit a certain number of games per month. Because if player A plays 5 games and plays them all early in the month and player b plays 5 games and plays them all late in themonth there should be no differnece between them. Its just the poor fool who plays 10 games throughout the whole month whos last 5 games start to have less and less value so it almost would be worth it at a certain number of games to save em and wait until the new month to submit em.

What this seems to be rewarding is consistent submissions. If you go gung ho on GalCiv and play 15 in one month and then disappear for 2 months you're worse off than the player who consistently submits 5 games per month over those same 3 months. Slow and steady always did win the race I guess.
[Message Edited]
[Message Edited]
Reply #19 from Stardock Forums Top
So it's not rewarding players who play the game a lot?? That seems kind of strange.
Reply #21 from Stardock Forums Top
Also means that at the start of each new month, overall scores will jump up very quickly and of course scores during the end of months will tend to not do much to the metaverse.
End of quote


Remember that the 'end of the month' in the metaverse is the date of your most recent submission, so the status of scores being at the beginning or end of the month tends to change with each submission.

The time to look for score jumps is just after the first couple of 'month anniversaries' of the release of version 1.05 as the first of its large scores enter new blocks.
Reply #22 from Stardock Forums Top
I wonder, are 1.05 and 1.09 considered in the same block?
Reply #23 from Stardock Forums Top
Blocks are 30 day periods determined by the time of the player's most recent submission. The full score block has all scores 0-30 days before last submission. The 90% block has all scores 31-60 days before last submission and so on.
Reply #24 from Stardock Forums Top
Oh, I misunderstood you, IceMole. I thought the blocks by versions.
Reply #25 from Stardock Forums Top
In light of this, I think players are dying for their scores to age, and get that 90%/80%/70% penalty.

Who would have thought, the more spread out your scores are the better.