[JCrowe]

MORE NEWS FROM THE FRONT …

… and it’s pretty ugly if your native language is based on one’s & zeros.

1.) ISLAND FATALITY SYNDROME

Problem:
AI incapable of developing itself when it begins the game on an island.

I’ve noted this before, so I’ll keep this brief – when a computer opponent starts the game on an island or small continent (under 13 territories), it gets ‘stuck’. The AI never invests in transportation or makes any attempt to move beyond its initial land mass. At most, the AI will build up basic land units until it reaches the limits of its resource capacity. Or it will build up a smaller army, leaving a budgetary surplus open. It then sits on the surplus, accruing a mighty hoard of cash … that it will never, ever use.


2.) “RANDOM” MAPS? NOT QUITE …

Problem:
Map Generator loves to drop computer opponents on islands, and human players get continents.

I’ve played a bunch of games now, and the map generator is demonstrating some predictable patterns, none of which are good for play. For one, it loves to drop computer opponents on island territories, even when vast continental arcs stretch the ‘globe’. In game after game, the majority of computer players in my setups have gone “Gilligan” for their three hour tour. In more than one game, ALL computer opponents were “Cast Away”. Which wouldn’t be so terrible, if not for Island Fatality Syndrome. With the AI unwilling to act or take initiative, human players basically win by default on day one.

On the flip side, I’ve never had to deal with being a passenger on the S.S. Minnow myself. Invariably, no matter what the setup, I, Human, always get planted on a nice, sweet, juicy continent. Not always the BIGGEST, but certainly one of the best. Again, not that I mind, but considering all the other issues, it just makes an unfair situation that much more lopsided. It would be interesting to have to play from ‘the other shoe’.

In fact, playing against the AI and the awkward predilections of the Map Generator makes the whole situation seem very reminiscent of an old “Twilight Zone” episode. You know the one – bank robber gets shot during a heist and ends up in the afterlife. He thinks it’s heaven, because the booze is free, the girls are easy, and every roll of the dice or hand of poker goes his way. He can’t lose! And then he realizes … he can’t lose. Sure, he can deliberately throw the game or lose his money, but it’s all his choice. It’s all within his power. And just as easily as it was lost, so it is claimed once again with half an absent thought.

That’s when the robber understands that he didn’t go to heaven ….


3.) “AIR” TODAY, GONE TOMORROW

Problem:
In only one game of a dozen did any AI team build air units, which are critical to game play / success.

Mystery of the disappearing air units. In the very first game I played, every AI in the field was pushing aerial artillery. One team even went so far as to invest nearly all of its cash in bombers (force of 10!), with just a smattering of tanks behind for tagging cleared-out turf. (Not a bad idea, if you actually apply yourself and attack someone – see issue #4 below.)

However, once past this game, the AI never again invested itself in air power. And I’m not just talking about bombers and jets – the apparent ‘injunction’ covered choppers and surveillance craft, too. (Not that there’s much use in WS for AWACS – you’ll “be there” next turn anyway, and it’s not like they have anything that can stop you …)

Only once, games later, did I see another AI air unit – a single ASW chopper that the computer built in a territory adjacent to a pair of boomers I had parked off his coast. I had been vaporizing the AI’s military with total impunity (conventional arms only) at my extreme leisure when this little beauty popped up on radar.

Ouch!

I could be in trouble!

… or not, as the AI flew the ASW chopper away from my subs, and straight into a nearby territory I was about to invade. Apparently, the chopper – the ASW chopper – was built with the intention of supporting his tanks and rocket artillery units in their grand attempt to deter my seven fighters and five bombers from charging into the capital. Mission: Failed. Dude – they build institutions for intellects of this level.


4.) DOVES VS. HAWKS

Problem:
The AI is a total “dove”; reluctant to attack (anything) and disinclined to push for new territory.

The AI lacks proper aggression. It almost never attacks, even in those few rare occasions when it might have the opportunity to do so; at most, it might move to reconquer turf you just seized from it and then vacated. More fundamentally, it fails to pursue territorial acquisition with any demonstrable rigor. I’ve had plenty of games where an AI team – totally unhindered by neutrals, other AI forces, or “me” – utterly failed to cinch up even a modest-sized continent. By the time I’ve gotten 24 territories under my control, AI X or AI team Y might still be plodding along with six … or seven. By the time I’m up to 43, they’re barely into double digits and still have oodles of open space to seize in their own backyard – and I’m not talking about island-jumping, either. Contiguous land-mass, dude. Two tank divisions could have covered the place in four turns … where are you?

Story is the same whether I add or redact neutral nations. AI doesn’t push hard enough, even when it has the units available to do the job. Did someone forget to cable? They do understand that there’s a war on, don’t they? Why the lackadaisical attitude?


5.) IF THE AI DID HAVE A BRAIN …

Just Speculating:
If the AI does get an upgrade, something might need to be done about –

i.) Map Wrap Around – not being able to “wrap around” the globe is limiting, and it also acts to unfairly protect players from attack. They’re free to ignore the “back door” and concentrate their forces “out front”.

ii.) Economy Needs an Upgrade – As I’ve suggested, cities ought to double a region’s cash. Otherwise, in the face of an “open” world map and competent opponents, players will have great trouble adequately defending their turf. The cash just simply isn’t there to provide adequate defenses. And I’m not talking about fortifying every territory you control, just building a few “rapid response forces” sufficient to counter any real threat is a strain under the current economic system. If not fixed, you might end up with a situation where the snake is constantly devouring its own tail – I seize your unprotected turf, then you seize mine, etc. etc. Perpetual war.

iii.) Tactical Combat Needs a fix – It’s a take all / lose all construct. There’s virtually no chance here for “Mr. In-Between”. And while human players currently have the upper hand, I can easily see how a good AI could completely reverse that scenario. Determining / planning “initiative” in WS is an art form. I mean, at least for me …. I can generally predict which unit is about to move next, especially in simple fights. But bigger battles with many unit classes or fortifications turn expectation into abject mystery.

Sure, stealth fighters usually ‘go first’, but how do you predict (in a complex battle) if the stealthie has enough “points” (?) to move into someone’s shooting range and then blast ‘em before they can return fire? I’ve had plenty of instances where I held high-initiative units back to prevent the opposing forces from getting the “jump” on them, only to find that my guys dominated the next round of initiative “dodge-ball” and could have wiped-out the other team if I had moved forward … of course, on those times that I have moved forward, expecting to hold the advantage, the tables have turned and the other side gets to go before I do.

But a good AI will be able to crunch all the calculations, predict the variables, and precisely time its moves to maximize its damage in ways that very, very few humans ever will. (Paging Mr. Kasparov …”) Then, human players will do all the dying and the AI will be able to march straight through to the doors of Rome with nary a pause in its digital step.

An “Axis & Allies” approach to combat resolution might be a better course.