Page 1 of 1

Retreat into combat

Posted: Thu Mar 11, 2010 11:56 pm
by squarian
1756 campaign, playing Austria vs. AI, now in mid-57.

One thing I'm seeing repeatedly and am somewhat disturbed by - armies either do not retreat from a region after losing a battle, or else they retreat in the wrong direction toward another enemy force. The result is repeated battles, often a series in the same turn between the same forces, with extraordinary casualties rates.

In any given battle the casualties seem about right, but the effect of three or four or five battles in the same turn can be devastating. For example, a force of about 20K under Henry of Prussia just fought five or six separate battles with various Austrians, bouncing back and forth between Korneburg and Wien. In each engagement the casualties were no more than about 10%, but the cumulative effect of this turn was about 50% casualties for both sides.

Shouldn't a force which has lost a battle retreat away from the enemy rather than back into the arms of a second enemy force? Maybe this is WAD, but it seems to produce some very odd results.

Posted: Fri Mar 12, 2010 3:15 am
by Marquee
Armies of the period should always always always retreat back on their supply line if at all possible. Only if that path is blocked should they consider a different route.

Posted: Fri Mar 12, 2010 9:38 am
by Pocus
I'll inspect that, there is already quite an hefty algorithm supposed to make retreats toward friendly lands, so there must be a bug somewhere...

Posted: Fri Mar 12, 2010 10:16 am
by Pocus
I can't use your save, I need backup1 (ie the turn before the retreat) to reenact the battle. You gave me backups 20 and 21.

Posted: Fri Mar 12, 2010 11:49 am
by Jarkko
It is logical that troops retreat towards supplies, but on the other hand it is suicidal. Something that is very visible in the Saxony scenario when you play the Imeprial side (no wonder it was so easy to overrun Saxony when I played Prussia :D ) -> Saxon troops simply will not retreat out from Leipzig area, they stay around and fight again and again until they are killed to the last man; if you try to be clever and move them to the harsh terrain south of Leipzig, and the troops get pushed out from there, they happily retreat to Leipzig even if it is under siege (and get then promptly slaughtered).

Posted: Fri Mar 12, 2010 2:40 pm
by squarian
Pocus wrote:I can't use your save, I need backup1 (ie the turn before the retreat) to reenact the battle. You gave me backups 20 and 21.


Oops - here you go.


It is logical that troops retreat towards supplies, but on the other hand it is suicidal....they happily retreat to Leipzig even if it is under siege (and get then promptly slaughtered).


This may very well be what I'm seeing in some cases - retreating toward supplies and into suicide. So if this isn't actually a bug in the code, maybe the parameters need to be reevaluated - toward supply unless that means an area occupied by enemy, in which case choose the next nearest, etc.

But it can't be the only explanation, because my savegame shows Henry of Prussia retreating from Korneburg to Wien (and he pulls off this brilliant maneuver several times). How can Wien be closer to Prussian supply than Korneburg to the north?

Finally - if this does need a fix (and I hope it does :D ), it might be a good idea to check all other AGE games for it also, since I've seen the same behaviour in WIA and ACW.

Posted: Fri Mar 12, 2010 2:53 pm
by squarian
Drat - I just realised, I played several more turns after posting my original message, so when I uploaded the savegame I gave you the most recent turns instead of the turns when the Wien-Korneburg battles happened. I think the turn in question was backup4 or backup5, but to be safe I'll post backup3 as well.

Posted: Fri Mar 12, 2010 3:33 pm
by Pocus
I did a fix for the next public beta patch along the same trend... I realized that the presence of enemy troops in the retreating region was not a repellent... Now this has been added as a parameter to consider.

Posted: Fri Mar 12, 2010 5:59 pm
by Longhairedlout
Hi Pocus, any idea when the next public beta will be released?

Thanks :)

Posted: Fri Mar 12, 2010 8:43 pm
by Pocus
It seems rather stable, so tomorrow it can go public (beta public) - but still stable. For an officially approved patch, there will be a delay of at least 2 weeks though.

Posted: Fri Mar 12, 2010 9:22 pm
by Longhairedlout
Cool, thanks Pocus :)

Posted: Sun Mar 28, 2010 6:44 pm
by squarian
1.01i, playing Austrian, 1756-63 campaign vs. AI:

Another odd retreat result:

Wutginau's force besieging Frankfurt-a-M, I attack and win, he retreats to enemy territory at Mainz and lays siege to it instead.