Auto resolve

[Loki]

NOTE: This was originally a reply to the Battles vs. Strategy thread, but not wanting to hijack it I thought better to post fresh.

Greetings and Salutations All

I'm a newbie here, this is my first post, so I'd first like to say thanks to all the hosts and moderators for this great forum. As has been noted on other threads, if you don't happen to have a close friend(s) that play Total War, then it is SO great to have a forum like this to speak with kindred spirits

Battles vs. Strategy

The longer I play MTW the more I'm getting into the stratagy. I must admit that when first playing the game, I was in a mad rush to get to the next battle and test my tactics. http://www.totalwar.org/forum/non-cgi/emoticons/wacko.gif

I've had great good fun and countless hours of enjoyment playing MTW, experimenting with various factions etc. My latest mad experiment is exercising my Sun Tzu, Musashi, and von Clausewitz, by not fighting ANY battles that I am not SURE to win. I am doing this by using auto resolve for EVERY conflict. This has been an enjoyable experiment but quite puzzling. Outcomes have not been consistent nor logical. (Yes I am accounting for the rank of the generals involved, troop valor etc. etc. As a matter of fact, I’ve become so mad I’ve started to keep a log book next to my keyboard to catalog ALL units anywhere near or within striking distance of my frontier.)

Mind you, the quirky “Auto resolve” hasn’t always gone against me…. There have been several times that I’ve been caught off guard and had to fight when outclassed or outnumbered.

I would be very interested if any of the old-timers here had ever heard any of the developers speak to this issue. What exact factors determine the combat power of an army? Or more specifically, what is the exact logic behind “Auto resolve”?

Thanks All
Loki

[Shamus]

It is my understanding that the AI looks at the units involved, their level of moral, armor, etc., as well as the level of the generals involved. Then the AI works out a series of percentages for the number of loses of each unit involved. The results can vary within a certain range of acceptable outcomes. This keeps things realistic in the sense that the same combination of units will not register the same outcome each and every time they meet, and at the same time ensures that there wont be many unrealistic “blowouts”.

That is my understanding of how it works, but please correct me if I am wrong, or if I missed something.

Welcome to the boards. http://www.totalwar.org/forum/non-cgi/emoticons/smile.gif

[Loki]

Thanks James (or should I call you Seamus if you REALLY want to be Gaelic?) http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif .

Knew about everything you replied with except the random factor part, (Kinda sounds like the infinite improbability drive&#33http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif http://www.totalwar.org/forum/non-cgi/emoticons/wink.gif

That seems to be the missing piece of the puzzle that explains the unexplainable

Kind Regards
Loki

[Brutal DLX]

I'm not sure about this one, but I recall one of the devs stating something about it a long time ago, perhaps a search can yield some answers.
From my experience, however, there appears to be no random factor, each units' combat strenght is examined and weighed against those of the other army, with all bonuses from generals and possible castle defences. Usually the results are reasonably congruent with what would happen if you played out the actual battle controlling your armies like the AI would, ie no super smart tactics, basically a clash of forces with some flanking. You can see that the inferior army consisting of peasants often has quite a lot more losses, or that spear heavy armies fare better against horses, all other factors being equal. The only inaccuracy that I have observed is, in my opinion, that heavy cavalry such as Royal Knights do get a bigger edge than they would normally have, thus you see the AI often attack with the king and his heirs and just a few bad troops and win on the autocalc or even forcing the enemy to withdraw without a fight.

[Cruelsader]

The autocalc results for storming strong castles were ridiculous in MTW 1.1 and 1.0. (do not know about v 2+ since I have not autoresolved a siege). If there were very few defenders, say 5 men, the attacker often lost only 1 or 2 men even when attacking without siege artillery. However, if you tried to do it yourself youd would always suffer much greater losses even with siege equipment.

[G_Love]

Quote[/b] (Cruelsader @ Dec. 09 2003,08:18)]The autocalc results for storming strong castles were ridiculous in MTW 1.1 and 1.0. (do not know about v 2+ since I have not autoresolved a siege). If there were very few defenders, say 5 men, the attacker often lost only 1 or 2 men even when attacking without siege artillery. However, if you tried to do it yourself youd would always suffer much greater losses even with siege equipment.
I remember reading a post hereabouts that this was because of a bug that caused those autocalc battles to not factor in castle defenses. So basically, it was 5 guys against your army on whatever the terrain for that province would normally be. Thus, almost no casualties for you in battle. I believe this was fixed for VI (2.0).

[Shamus]

Perhaps someone with more understanding of the game could help clear this up. The impression I had gotten from what I read, was not that percentage chances of how each unit fared was over a wide range. Rather, it was over a short range. For example, a unit of urban militia with no armor and a general of one star, will always beat a unit of peasants with no armor and a general of one star. However, the number of surviving urban militia will not be exactly the same each time. This is my understanding at any rate.

[Gregoshi]

Greetings Loki. http://www.totalwar.org/forum/non-cgi/emoticons/wave.gif Welcome to the Entrance Hall.

I vaguely recall one of the developers discussing the auto resolve, but that may have been back before we moved to these new boards.