AMAI 1.9 bug?

Single player works fine all the time,but everytime i start a multiplayer game it boots the human player.I really love to pratice(AT) on AMAI,and the lastest and greatest so far is this release(1.9).

also ive noticed the major lag at the begining of the game now.

great job on the AMAI!!!!
I used to beat AMAI normal on a reg. basis,now with hero pot abuse and the hero saving TP's, im lucky to win 1 out of 10.

I'd imagine the lag is caused by the 'scanning' of the map, because as far as I know it moves a unit over the entire map and checks to see different effects. This will un-doubtably cause lag, but AMAI hasn't caused that much for me.

I, too, suffer from a short severe lag at about 5 seconds from the start of the game. And I play 1vs1 on a local machine (AMD 1.53 GHz, 512 MB RAM).

You are right, the severe lag in the beginning is caused by the scanning. I could spread that over time but in my tests that made playing harder during the whole time it scans.

As for multiplayer games, there is an issue that the Blizzard.j is not loaded when it is the first map you play so that could cause a desync.
So my question is, was that with or without the Commander and when exactly did they get booted? What where the races used?

I suggest that you inform human players just before the lag, e.g. as follows:

"Initializing AMAI, please wait. This may take a few seconds."

It's good to inform users that the lag is an intended functionality.

Thats a good idea. I will put that in the next version.

I tested it at a lan it didnt work with or without the commander.
I hosted played UD, my ally was human, he got the boot.
We tried it with all the races, same result.

btw , we loaded up 1.8 and it played fine.

Hmm, that sounds strange. When exactly does he get booted and is it always the one not hosting that gets booted?
Was it the XMas-Release or 1.9 Release 4 ?
TFT or RoC?
Was the Commander just deactivated or not included in the map?

Just curious: What exactly are you scanning for?

I am scanning where you can walk and where you can't. By this I get a pathing map on which I can then run pathing algorithms to determine if a point is reachable from another by foot and how far away it is.