RE: How is the AI?

Matrix Games Forums

Forums Register Login Photo Gallery Member List Search Calendars FAQ

My Profile Inbox Address Book My Subscription My Forums Log Out

RE: How is the AI?

View related threads: (in this forum | in all forums)

Logged in as: Guest

Users viewing this topic: none

Printable Version

All Forums >> [New Releases from Matrix Games] >> Strategic Command Series >> Strategic Command WWII War in Europe >> RE: How is the AI?

Page: << < prev 1 2 [3]

Message

<< Older Topic Newer Topic >>

RE: How is the AI? - 11/26/2016 7:40:51 PM

itkotw2000

Posts: 30
Joined: 8/27/2005
Status: offline

I agree with Polonthi, the AI is greatest on the small scale. What this AI does best is manage single attacks and shuffle units. I was amazed at how well the AI shuffles units in north Africa (very narrow frontage). The strategic AI will not surprise you with its creativity, but on the other hand it doesn't make surprise you with massive mistakes either. So, at least the strategic AI is good in that sense. Like a lot of people have said, the AI relies on scripts and some scripts have been tweaked more than others. When I think of "good" AI programing, I think of how well the AI does things on its own, not things like "invade Spain 10% of the time". So when I see the AI shuffle units in an attack, I say "wow".

Strategically, the AI can beat you, but not because it outsmarts you. My opinion comes from only playing axis and only playing to early 1943 (I always restart to try something new, a.k.a. "learned"). Certainly in the early war, AI scripting dominates the AI Allies strategic choices. Like the French army sitting on the Belgium borders waiting for the Germans or the Russian troops the appear on the border when you declare war. The Russian border being empty shocked me the first time I played. I even went back and counted units and hexes. The eastern front border is 51 hexes long and only had 6 Russian units on it.

The Russian army is not scripted to put up a fighting retreat or counter attack, it is only scripted to occupy cities in the probable path of the German advance. Towns/Cities and how they work in this game do the AI thinking for it, you could say. Strategically, the AI only knows to attack from one town to another. You will NEVER see the AI attack across an empty area as part of a big push. The AI will surround your unit in a town, the AI does not literally advance straight at the town hex. Or the AI might send a unit to try to cut your supply lines. This is one of the areas the AI makes mistakes. It will rush a unit or two to try to cut supply, but those units are easily surrounded.

There are some game mechanics that make it seem like the AI is making good moves, but it is hard to tell if the AI was made to attack considering the game mechanics or whether the game mechanics were added/changed to help the strategic AI function. I will list a few mechanics that help the AI. I mentioned above that the AI will never make a large attack in an open area, but you will notice there are no open areas. The map is filled with towns and cities just so that the AI doesn't have to deal with open areas (this compounds when added to other mechanics). Also, every town is a supply center and produces some supply. Even towns that are totally cut off produce supply (3 supply points). A friendly occupied town supplying your troops has average supply of 5. So, your units 2 hexes from a friendly town have the same supply as troops in a cut off town. Another point, most of these numerous towns/cities are conveniently linked by rail, creating a nice web for the AI to follow (this gives easy options to the AI). And finally, units that are cutoff in towns can rebuild some their strength and reinforce.

You could reply, "hey, the map has a lot of cities, because they were really there and important. And troops that are cut off could forage for supplies. And Europe had lots of rail lines, that's why they are so connected. And maybe the units that are surrounded rebuild their strength by gathering stray troops from other routed units". When you consider the scale of the units in this game, armies and corps, then it makes a lot of these things less plausible. Imagine an army (many many divisions, with all of their support units, trucks, artillery, and tanks) in/around a town. How many cities on the game map could plausibly support an army when cut off? How many scattered troops and equipment would have to be gathered to increase an army's strength by even 1/10th (1 strength point)? How many rail lines in the 1940's could support multiple armies down a single link, every link marked on the map?(every railroad on the map can support an infinite number of troops).

It is interesting how many people saw Polonthi's post and saw a problem in diplomacy. If he takes all of his resources that would go toward a navy and uses it for diplomacy, is it that bad? I read his post and though, "yeah, I wish the navy was more useful". Neglecting the navy should have some impact and the player should feel the navy is useful. Diplomacy in general is "weird" because so much of the MPPs are pooled together.

(in reply to The Land)

Post #: 61

RE: How is the AI? - 11/26/2016 8:08:30 PM

Hartmann

Posts: 888
Joined: 11/28/2000
Status: offline

The diplo thing is really an issue if the AI does not counter human moves in the diplo game.

About the navy, I think it actually is very useful in this game. Subs attacking convoys hurt MPPs ALOT! So to counter that, the Allies need lots of destroyers. The Italians have to wipe the Royal navy from the med if they want to achieve something in Malta and Africa, and the British have to do the same with the Italians. The Germans need to get the naval upper hand in the North Sea if they want to do Sealion.

Granted, you CAN win the game as the Germans without a navy by just beating the Soviets plus winning the diplo game to get Spain on your side, take Gibraltar etc. That was always one of the alternate history ways to do things - even in Clash of Steel. This should remain possible in principle, but here we are back to the problem that the allied AI has to take diplomacy seriously.

About the AI trying to cut of supplies: I like that it does that, but it just is not as good at it as a human player can be. If one player tries to cut off the enemy, he usually creates a possibility to get cut off himself instead. And the AI indeed has problems seeing who would get the upper hand in the "encirclement game".

(in reply to itkotw2000)

Post #: 62

RE: How is the AI? - 11/26/2016 9:40:12 PM

xwormwood

Posts: 1149
Joined: 8/28/2000
From: Bremen, Germany
Status: offline

The AI has often cut of my advanced units which received the very same order from me (storm into Russia, cut through rouds and rails. On many games you can push the AI with such a behaviour, as it tries to secure the hinterland (moving lots of units to place them into open cities and towns. But it is different in this game. Here the AI will counter your move, at least as long as it has enough units. So there is no real weakness which I could confirm. At least not in this part.
But the AI should always have enough MPPS to be able to get enough units to stay such a dangerous opponent, as it can't act completely like a human without looking "dumb" or "overly aggressive".

_____________________________

"You will be dead, so long as you refuse to die" (George MacDonald)

(in reply to Hartmann)

Post #: 63

RE: How is the AI? - 12/1/2016 4:39:26 AM

EdwinP

Posts: 179
Joined: 2/20/2004
Status: offline

quote:

ORIGINAL: Hubert Cater

Agreed as we'll look into having the Allied AI be a bit more reactive to this type of Diplomacy.

Perhaps add diplomacy scripts that would trigger based on a country's activation %d date. Much of this is doable in the current game engine.

Example for UK: If Spain is Pro-Axis 60% or greater then purchase Spanish diplomacy chit. (Max 2)

The one criteria that needs to be added is a check for activation % less than a specific level (as opposed to a minimum activation %) or in a range of activation %.

Example for UK: If USSR war readiness is less than 60% in Dec 1940 then purchase USSR diplomacy chit (Max 2)
Example for UK: If USA war readiness is (1% to 50%) in Dec 1940 then purchase USA diplomacy chit (Max 2)

Example:

; #VARIABLE_CONDITION= Under what variable conditions will this event occur
; Format: country_id [political_alignment][activation_range][min_activation%][surrendered_flag]

Example: 112 [2] [1,50] [0] [0] - Triggers if activation % is between 1 and 50. Does not trigger if over 50.
Example: 112 [2] [1,100][80][0] - Triggers if activation % is 80 or higher.

< Message edited by EdwinP -- 12/1/2016 4:05:32 PM >

(in reply to Hubert Cater)

Post #: 64