REFUND!

seriously, could i at least get back my two buy-in points that were not calculated in the final scoring?

Well, think of it the other way trip: this happens to everyone else as well.

From objects/game.php: Function pointsPerSupplyCenter

totalSCNo is the number of all SCs owned by players not in CD.

return round((float)$this->pot / (float)$totalSCNo)

So sure enough, it calculates a points-per-supply-center, although not necessarily pot/34

So in a 12 bet game with no CDs, only 68 points will be distributed, 16 will go *poof*

In a 8 point game, 68 points will also be distributed. Only 56 points were bet.

i think trip deserves his points.... not to mention that he's found a major bug that affects us all...

thanks trip!

i know figs, i'll get over it. just not really thrilled with the outcome right now.

anybody want to start of series of 8 pt games? we can 7 way draw and turn a profit.

Lol. That wouldn't last :)

I wouldn't call it a bug. It's a design decision with perhaps unintended consequences. I'm guessing they never actually sat down and looked at the percentages for low-point games.

The extremes:
With a 7-point buy-in, the players lose 30% of the bet.
with an 8-point buy-in, the players gain 21% of the bet.

The ONLY bets where the payout on a non-CD game is the same as the buy-in are exact multiples of 34. Makes sense.

LoA - its not a newly discovered bug, but unfortunately it hasn't been dealt with yet. Its not publicised to stop people trying to exploit it. I'll get in contact with Kestas and ask him to make that rounding into a floor, so the system can't be exploited before a real improvement is made

Draws are different. From function setDrawn:

$winnings = round($this->pot / $players);

It's no longer points per supply center. Now it's points per players. Since every pot at the beginning of the game is divisible by 7, all points are awarded in that case.

Now, if you can knock one player out then draw, you can gain a few points, but no more than 3:

On a 15 point game, there are 105 total points. When divided by 6, everyone gets 18 points for a total of 108 points.

Making it a floor (actually, an INT) will make it unexploitable, but also guarantee that as a group we can only LOSE points.

Now, that has to be balanced against the point inflation caused by the immediate bump to 100 points after losing a game. We get a net inflation, considering the rate of newbies joining.

jbalcorn

i think, by definition, that it's a bug.

My suggestion of a floor was as a short-term measure as Kestas seems very busy at the moment, so changing like that (yes fair point to use typecasting though!) would be very quick to add.
It would be far better to round at the last minute.
If you have better suggestions please do make them on the dev forum though!

But still, avoid games with bets of 7,6 and 12 points. 30%, 19% and 19% point loss. The next worse is an 11 point bet with a net loss of 11.6%. Those are the only ones above 10%.

Until they change the round to a floor, 8, 17, 13 and 22 point bets are the best for low-point with high possible reward.

Talk about a house advantage, though. I think I'm going to take the code offshore and set up a play-for-money site! :-)

Yeah. Looking at the code, it's going to take some significant refactoring to fix this. It's not a bug, I think Kestas was very aware of what he was doing. I think it's a design decision. Since there's no actual money involved, and he gives away 100 points over and over again to everyone, it's not that it's unfair.

However, DO NOT play this game for money.

figs, thanks for your help. i appreciate it.

has this been fixed yet? i think the system owes me more than 250 points. i have done the calculations myself

i would say it's a major bug

it's a 'design decision' that is obviously very wrong and completely unintended. i.e. it's a bug.

Why? It's applied to everyone and is roughly even. Moreover the common values of 100 etc are 'wins' and as such you probably owe the system.

Alamothe - this "investigation" only really got going in the last few hours, so no it hasn't changed. Be fair - Kestas is not only very busy but I would also bet he's currently fast asleep. Remember world timezones!

it's called a bug because it makes the system behave in an unintended, and usually incorrect fashion.

that's what a bug is.

for coders i mean.

i'm sure kestas will fix it or do something about it... so not to worry...

fig, why can't that function return float, and then do the rounding elsewhere?

Actually a 7 point buy-in game with someone buying a 2 point CD to put the pot total at 51 gives the biggest extra percentage back with 68 points going out for a "profit" of 33.3%.

In any case, if you stick to multiples of 5 any loss would be pretty minor. Except for 50. But even that is just a loss of 10 points out of 350.

BTW i'm not saying it must be fixed in the next few hours :-) we have been playing with this bug for decades, right?

Since the introduction of the points system yes

It could return a float yes, and that would be the prefered solution.
Using int rather than round would be a good stop-gap method to prevent deliberate exploitation