Author Topic: strategic use of units  (Read 1218 times)

discodowney

  • Guest
strategic use of units
« on: 16 December 2010, 18:07:08 »
Hi,

I realise what im about to ask can be figured out playing the game enuf, but im running really low on time so any help ye could give me would be great.

Basically, for the tech side, im wondering if people could give an idea of when they think the different units would be useful. So catapults would be useful when there are, say, more than 3 towers for the ai to deal with. But when should the AI choose Horsemen over Guards. This is for both offense and defense. What would be an optimal number of each unit in an attacking force?

Cheers

wciow

  • Behemoth
  • *******
  • Posts: 968
    • View Profile
Re: strategic use of units
« Reply #1 on: 16 December 2010, 20:49:48 »
Ok heres my take on how the tech AI should play based roughly on my own strategy.

Early game:

Tech players should concentrate on defense in the beginning. Currently the AI is quite agressive early on, this can lead to easy defeats against magic since any good magic player will put all their effort into churning out as many battlemages as possible early then allow a tech force to be beaten using battlemages massed behind the golem and then counter attack to win. This tactic is very good vs tech.

Techs strength early on lays in it archers, so the AIs main goal throughout the early/middle game is to mass produce archers for defence. A good tech force should consist of mainly archers with a few swordsmen to act as a screen/meatsheild whilst the archers rain on the enemy. Towards the end of this phase the AI should be progressing towards producing guards rather than swordmen and also working towards getting horsemen.

In terms of upgrades the AI should upgrade training_field asap, i usually build two barracks. Once the second barracks is complete it researches this upgrade straight away whilst the initial barracks continues to train archers. The AI should also research Peircing weapons before blade weapons as this benefits archers (the bulk of your troops) rather than your meatsheild infantry.

Middle game:

During the middle game tech should continue to keep a defensive force of archers/guards in the base whilst also stockpiling horsemen. Judging the right balance between attack and defense can be difficult here since horsemen are quite expensive. The AI should err on the side of caution here. Once the AI has a few horsemen it should split its archers and send half of them with the horsemen as an assualt force. Whilst the enemy is fighting these troops quickly build up the defensive archers to the previous level and repeat.

During this phase the AI should be looking to research the armour upgrades and produce a technodrome.

This is also the right time for the AI to expand to secondary bases as gold may be getting short. To do this it should send a scouting force of 2-3 archers to the nearest gold position. Once there, position them as scouts at a distance from the gold and send workers to build the castle. If any attacks are detected retreat the workers slightly and send whatever horsemen you have plus half of your archers. Once engaged send the workers back to help save the castle by fixing it. Whilst it is fairly easy for a human player to work out whether an attack is a diversion or not the AI must be very careful not to send a full assualt force for 1 or 2 wandering enemy scouts.

End game:

Once Tech has a technodrome (or two) attacking waves should consist of screening waves of horsemen/archers send to kill enemy infantry and cause disruption followed closely by heavy waves of battlemachines, catapults and airships to destroy enemy buildings. At this point if an attack is looking promising (e.g. more than x enemy buildings destroyed) then any guard force left around the base can be thrown in to hopefully provide the final push to victory. However this can lead to the enemy snatching a surprise win if done at the wrong time. Again this is much more easily judged by a human than an AI player.

Lastly I wanted to say that the AI should give up on creating ornithopters. Currently it does not use them properly. The only real use seems to be as screens for airships, giving enemy ranged units more targets to spread damage. Unfortunately the AI prefers to use them either as lone scouts or suicidally attack your base with a fleet of half a dozen ornithopters which are instantly shot down by any archers/battlemages you have.
Check out my new Goblin faction - https://forum.megaglest.org/index.php?topic=9658.0

Yggdrasil

  • Local Moderator
  • Ornithopter
  • ********
  • Posts: 408
    • View Profile
Re: strategic use of units
« Reply #2 on: 19 December 2010, 16:59:23 »
One minor thing which should improve the AI quite nicely is to move their attack units in groups and attack together. And try to attack a position from different direction, so the player can't just build towers at one side of his base. These two things are a bit difficult to get right i have to admit but it should make matches more interesting.