It's random with a slight weight so that you're less likely to play a country you've recently played.

Yup. What they said above. Although I got Britain in two live classic games in a row yesterday and a Britain on the modern map as well, also yesterday. :P

Who is jhoffer007?

Ok thanx for the info.
Oh and I am jhoffer007....you will be seeing that name alot....heheheh

¿IS HE KNEW?

What abgemacht said, but also consider that there are six other players on the board who also have weighted random countries, so this also affects the selection.

And britian in modern is not the same country (afaik) as england in classic.

i think you get a random country weighted only by the countries you've recently played on that specific map.

Also, the weighting is really "recently completed playing" so if you are currently underweighted in a particular country, your odds go up of getting that country in several games running concurrently.

Actually Bob your country chance weightings only come into play when a game starts, this is the only time country chance weightings are read from or written to; nothing happens when you leave a game.

To the original poster if you want to be a bit more specific each time you are a country the chance of you being that country next time is halved (there's a bit more detail but that's the nut of it)

@kestas, am i right in assuming it is weigthed per variant, so playing england in classic has no effect on the chances of playing britian in modern?

"Oh and I am jhoffer007....you will be seeing that name a lot....heheheh"

What an announcement ...... I'm a little shaken, but not stirred

@kestasjk ... I was not clear. The country weightings (as I understand it) are computed when the game starts, but only completed games are included. I thought I saw reference to that once, but it also matches my experience. Several times when I have joined 3-4 games between two "finish times" I have had multiple "duplicate" assignments. Over half the game I am currently in, for instance, have the same country assignment, and it is one I had not played previously for a long time.

I thought country assignment weighting was only done on the Classic map and not on any other variants. But I certainly could be wrong.

ok, i took a look at the db on my local version, and there are 7 variables only associated with each user - chance of playing the 7 countries in classic... but i can't see where they are used. Pregame.php (line:41) seems to use a shuffle to randomize the users for assignment, but without taking into account the chance, i might be looking at old code... also confused...

@ora, now that you mention that, I possibly remember kestas adding variant variables with a new db table so that the weighting could be done by variant in the new table as well as variants being able to do other things with the new table, and those 7 variables are no longer used. Standard disclaimer about being wrong.

ah-ha, well i can't find such a thing, but i may have an old version installed.

Hi VB, there's no distinction between completed and uncompleted games; the moment you're assigned to a game your chance of next being assigned is halved. You can always get unlucky and have a bunch of games in a row, but the code is applied when the game starts.

Also it is per variant, there's a new variant data table introduced a while back which allows per variant, per game / per user / per phase etc data to be stored in a flexible way, so that variants don't need their own data structures or to hack existing data structures. Not sure how widely it's used except for country allocation

So, kestas, does that mean that if a game starts, and I draw England, but then that game is cancelled in 1901 - that still means I am somewhat less likely to draw England in my next game?

found the code, it's actually way more complicated than i assumed. Because every player has a different chance of getting a particular country, this means that your chances depends on the other players in your game - which is messy... who gets to pick their country first? (in webdip's case we pick the person with the greatest deviation from the average 1/7) I might have looked at it the other way around - what is the chance of each player being england... randomly pick one based on the chance weightings... but i think that might be entirely equivalent.

@Jamie, i think that is exactly what that means.

Thanks, kestas -- nice to have the straight story.

@ora -- that approach is the only one that really works logically ... looks like a well-written algorithm to me.

@VitrualBob, i can imagine two equally logical approaches, one the used approach - take each player and pick a country out of a hat at random (starting with the player who has the biggest divergence from the norm), rinse and repeat

The second, take a country and randomly assign a player (based on their 'chance' variables) rinse and repeat. Would this result in the same results? possibly. I'd have to simulate both pieces of code, because my logic just can't tell the difference...