Looks like you are correct. The game ended before the retreat phase, which is incorrect.

I'm fairly sure that this is a known issue

It is a known issue but is a rather rare occurence so isn't as high up on the dev list as other items.

Cool, I was looking for an example of this. It is a well known issue. Really like a "house rule" of this website since it hasn't been addressed.

PresEden: it's not supposed to work like that. The rules state that the game should end after a fall turn, which includes the fall retreats and builds.

really? that's crazy, i didn't think it was an issue. You can calculate the win at the start of hte adjustment phase instead of the end of the fall move.

I didn't htink this happened at all...

@ orthaic: It's not crazy at all. You should not calculate the win at the start of the adjustment phase. Retreats are part of the turn.

In my experience this error only happens in Ancient Med games. I think the classic game handles this properly.

Jammiet99uk, no, it happens with any game.

This is a known issue. I would hope it's very high on the to-do list, as it essentially means the adjudicator is not faithful to the rules.

if it only happens in ancient med (as i think someone would have noticed in classic already and kicked up a fuss) then i guess the variant has some issues of it's own.

@Alderian, are you sure? i haven't seen it, but then i may be rare enough...

It happens in all of the variants. The adjudicator checks for the win directly after the fall moves before retreats

And it happens in all version of the classic map as well. It's just the way the adjudicator works. It really is rather rare though, at least in the classic map.

The quick fix is to wait until after the retreat phase. But then every game in which it isn't necessary is potentially extended by one phase. Personally I wouldn't mind, a bit more time to talk EOG stuff.

After that quick fix though there are a few other things that would be nice. Check whether someone is at 0 centers but not yet marked as eliminated and mark them so.

And before the retreat phase, if someone is at 18 or more centers, and if no one can retreat into a supply center, then skip the retreat phase and proceed to the winning conditions.

Add a two phase length timer after the game is completed in which new messages will still cause the new message tab to appear at the top of the page.

But, we get what we pay for, and frankly, I think it has been an amazing value. Thanks Kestas!

Yes, it happened to me too recently in a league game. It wouldn't have affected the outcome (I still would have lost) but I would have been able to retreat to a vacant supply center and upped my ending SC center count. Which is a tie breaker in the leagues.

fortunately in the league those running the game can manually adjust your end of game SC count...

@Alderian - yes we get what we pay for, now feel free to bump the donations thread!

I'm finished my current project so i can actually look at the source code for this adjudication...

My bad, I did not realise this was a problem in all variants.

If a mod / developer / kestas is reading, can you confirm this is on the 'to do' list? I agree with the others that it is a serious issue.

hmm, it seems that it works out if builds are needed or not - so you can't do this in the build phase... cause there may not be a build phase...

but wait, you can only win in the build phase. so that's alright yeah?

http://forum.webdiplomacy.net/viewtopic.php?f=3&t=448

Orathaic, one catch with that is while the winner obviously must have more centers than units, if the winners home centers are all filled, and if whoever lost a center to the winner had their unit disbanded due to no retreats possible, then it may not think a build phase is needed since no one could build or disband.