1 unit per hex: failed experiment

[FatCatAttack]

You made an analogy that worked to the opposite of what you wanted to prove, and now try to save it by redefining the context. I don't think this is a worthwhile venue of discussion, since obviously every failed analogy in every discussion can be "saved" this way. If you feel that pulling Persian catapults out of your hat adds something worthwhile to your original argument, fine, I won't debate it.

Again you are cherry picking now with selective quoting. We have to talk about catapults because we are talking about the general progression of the game. You are doing some projection here as you redefined the context from the beginning by crafting essentially a scenario type of situation trying to recreate the exact battle. But the thing is you are going backwards. My analogy is in terms of many kinds of Thermopylaes. Meaning we are playing the game normally from the beginning and perhaps we end up with a small empire and have to fight a larger empire. We are creating a Thermopylae situation dynamically to leverage our smaller force against a larger one. Perhaps I should have been more general to avoid nitpicking. Live and learn.

Who told you that Civ4 tried to "mitigate" SoDs? If the Civ4 design team had wanted to do that, then there'd be lots of possibilities (stronger stack counters like collateral damage, give units in a stack negative modifiers, implement higher logistics costs for large stacks, implement a chance that a non-optimal defender is chosen which increases with stack size, etc.). Mitigating stacks is easy. The problem is whether the AI can grasp the rules.

Civ4 chose the route to make large stacks the best option most of the time, and Soren wrote an AI that coped with them pretty well (assemble stacks, move and attack with them in a coordinated fashion, transport stacks across oceans, etc.). It's a design decision that paid off with an AI that actually poses a challenge for new players. Calling it a "failed experiment" by supposing that the devs wanted to mitigate a feature that you don't like is a bit of a stretch, imho.

Conversely, the design decisions for Civ5 were apparently to have a more tactical combat system and write the AI as an afterthought. This only way the current state of the implementation couldn't be seen as "failed" is when we assume that the devs didn't even care about making the AI competitive. But that's in contradiction to everything they said before.

It's interesting that Civ4's lead designer was also its AI programmer, which means that Soren could design the game's rules so that the AI could have a decent grasp on them. In Civ5, the AI designer is just one of many guys working under the lead designer, and apparently either he didn't object enough to a rules system that would be extremely difficult to teach to an AI, or wasn't listen to, or overestimated his abilities and agreed to it.

There's a dev who posts on the Somethingawful forums who wrote in a Megathread on Civ IV that they basically threw their hands up and accepted SoD's as an inevitability by BTS. That's why the protective trait came into being mostly to help the turtling computer and that was a failure as it screwed over quite a few leaders. Poor Toku.... The XP system is something that lends individuality to units. The SoD takes away that individuality . There's no synergy there. While some stacking was obviously meant to happen like a Spearmen protecting a unit from cavalry the mass consolidation of armed forces into a single tile was something they hadn't considered. Collateral was an attempt to fight the stacks but it just made it worse as your stack had to get even fatter to protect itself. Like some horrifying demonic cyst everytime they poked at the stack problem is just expanded the stacks even more.

Civ IV AI couldn't handle stacks. In fact the playability of the game ironically relies on the fact that the AI couldn't handle stacks. As I said unless you massively overpowered them equivalent forces would forever be deadlocked in weird little dance waiting for someone to take the plunge and move first. The AI isn't where the challenge is. The challenge is within the player themselves. Did you do this correctly? Did you optimize here? Did you prioritize the right tech etc.? It's a game that encourages rules mastery and metagaming. Civ V is a response to that. It tries to make everything develop naturally and dynamically and that's reflected everywhere in its design from the way it limits expansion to how units fight by plugging away at each other for multiple rounds to the social policy system. If they didn't consider Civ IV a failed experiment why change so much in response in Civ V?

what people do to defend their opinion.

Oke, i go along your way. I "imagine" that that cannon, does not represent 100% cannon regiment, but also some additional forces. Cool, nothing wrong with that.

But wait; lets take a Knight, for example. Following your path, i should imagine that that Knight is not made of 100% Knights, but also some additional forces. How come then, that those additional forces, can keep up with the Knight, who, apperently, travel much faster.

You see, i try hard. But i smell some inconsistanties in your reasoning. Now, whatever you answer will be; i bet that i have to imagine alot more

hoe about those great "exp upgrades" you can give your units. You have to imagine alot here also.
Now remember, the dev's have implemented "a PG style of making war" here. And may i add; poorly executed.
Oke, the horsemen and it's "exp upgrades". One of them makes your Horsemen runs faster (=more hex). Why ?
Did the horses get some new diet? Macnoodles ? And there are more "unrealist" upgrades. Very Gamey.

And you know what, it all would not matter that much; if only people didn't make so much fuzz about politics!
Cause that's where people DO complain about realism alot. If only they were so critical with the units they wage war with; i would be a happier man.

PG have updates too, but they were much closer to reality (until PG 3 came out, with indeed, also over the top goodies).

It was not unusual for infantry forces to make use of horses to go someplace but not necessarily fight mounted. I'd visualize a mounted unit in Civ is a unit of force that stresses mobility and shock. Longbowmen in the hundred years war who rode mounted to terrorize French villagers were not true cavalry but did so to move and leave quickly.

 

[Kiwi_Mark_LFC]

So let me ask this. How do a million different wargames do no stacking just fine with a competent AI? How is it that it's been done a million times before with no problem and suddenly now it's impossible to program an AI without a big stack? Seems like it's a giant crock to me.

If you have bad AI, then it's because it was programmed poorly, not because of a system that's been around forever.

This has been explained a billion times and is fairly obvious to anyone who understands anything about computer wargaming...they play on maps where 1 tile = only a few metres...not a few hundred kilometres...wargames are almost always about tactical combat and thus 1 unit per tile (which even then isnt entirely true) works...however CIV is a strategic game and thus should limit itself to strategic combat...and thus should allow multiple units per tile...

Moderator Action: <snip> Please do not use foul language, inappropriate comments removed.
Please read the forum rules: http://forums.civfanatics.com/showthread.php?t=422889

 

[jpinard]

Kiwi, please re-read my post http://forums.civfanatics.com/showpost.php?p=9770177&postcount=397
Hexes are just fine, limited units per hex are fine.

...and please take all those nasty things about Jon out of your post. That's despicable beyond belief.

 

[Psyringe]

Again you are cherry picking now with selective quoting.

I don't think my quoting changed the meaning of anything you wrote. If I quoted "selectively", then to clarify which aspect of your argument I was talking about, not to warp your words. I prefer this style as more efficient than full text quotes, especially when the quoted passage can easily be accessed by just clicking on the link to the quoted post. I do agree that I was "cherry picking" by addressing only parts of your argument, but would like to point out that you're doing exactly the same. Probably has to do with the complexity of the discussion.

There's a dev who posts on the Somethingawful forums who wrote in a Megathread on Civ IV that they basically threw their hands up and accepted SoD's as an inevitability by BTS. That's why the protective trait came into being mostly to help the turtling computer

Since that contradicts other info I have, I'd appreciate a link to the respective post.

Civ IV AI couldn't handle stacks. In fact the playability of the game ironically relies on the fact that the AI couldn't handle stacks.

I honestly don't see what you mean. I look at Civ3 and see an AI that sends a flood of units across your territory one by one. I look at SMAC and see an AI that sends units against your cities piecemeal. I look at Civ4 and see an AI that assembles stacks, monitors their composition, moves them in a coordinated way, uses one type of unit to prepare the field for others (siege removing defenses), evaluates the attack chances and priorities of the units contained in them before committing to an attack order, and even builds stacks of naval transports (complete with escorts) to transport stacks of land units around. I don't see any merit in your assessment that the "AI couldn't handle stacks", which you didn't even bother to flesh out in any way expect by a vague reference to a hypothetical "deadlock" in the rare event of two evenly matched stacks.

If they didn't consider Civ IV a failed experiment why change so much in response in Civ V?

Because, as Jon Shafer repeatedly said when asked about the early stages of the development process, "Civ4 is still a pretty good game." And it's hard to top. For the people who like the kind of gameplay that Civ4 has to offer, it would be very hard to create a successor that they like even more. Using Civ4:BTS as a base and building on top of it was also out of the question, since Civ4 was pretty confusing to many new players already, and adding more complexity on it would further limit the customer base. Hence, the only economically viable tactic is to do something that's still Civ, but different enough to tap into new target groups.

Civ4 is not a "failed experiment" and I see no reason why Firaxis should see a massively successful game as such. I think the opposite is true - Civ4 is/was too good to not pose a problem for its sequel. This isn't a new concept at all btw, although it's much better known in the movie/TV industry.

 

[Jediron]

Originally Posted by FatCatAttack
It was not unusual for infantry forces to make use of horses to go someplace but not necessarily fight mounted. I'd visualize a mounted unit in Civ is a unit of force that stresses mobility and shock. Longbowmen in the hundred years war who rode mounted to terrorize French villagers were not true cavalry but did so to move and leave quickly.

Let me tell you that it was quite unusual. Horses in those days were mostly driven by nobility and upperclass. Horses didn't grow on trees, you know.
But never mind, i can see where this is going to. Against such logic, there is no reasoning.

 

[FatCatAttack]

I don't think my quoting changed the meaning of anything you wrote. If I quoted "selectively", then to clarify which aspect of your argument I was talking about, not to warp your words. I prefer this style as more efficient than full text quotes, especially when the quoted passage can easily be accessed by just clicking on the link to the quoted post. I do agree that I was "cherry picking" by addressing only parts of your argument, but would like to point out that you're doing exactly the same. Probably has to do with the complexity of the discussion.

You change the meaning by framing my analogy in a context you designed rather than what I meant. Then you implied that you had "nailed" my argument. That is a very passive aggressive strawman IMO. Yes the discussion is complicated but if I leave anything out feel free to point it out and I'll address it.

Since that contradicts other info I have, I'd appreciate a link to the respective post.

http://forums.somethingawful.com/showthread.php?threadid=3137775&userid=0&perpage=40&pagenumber=1

SlightlyMadman is the user name. Afraid you are going to have to control F through 116 pages but it's discussed IIRC.

I honestly don't see what you mean. I look at Civ3 and see an AI that sends a flood of units across your territory one by one. I look at SMAC and see an AI that sends units against your cities piecemeal. I look at Civ4 and see an AI that assembles stacks, monitors their composition, moves them in a coordinated way, uses one type of unit to prepare the field for others (siege removing defenses), evaluates the attack chances and priorities of the units contained in them before committing to an attack order, and even builds stacks of naval transports (complete with escorts) to transport stacks of land units around. I don't see any merit in your assessment that the "AI couldn't handle stacks", which you didn't even bother to flesh out in any way expect by a vague reference to a hypothetical "deadlock" in the rare event of two evenly matched stacks.

That it's a rare event is in itself proof that the AI cannot handle stacks because it's not building enough of or the right kind of military composition. The AI wastes a lot of hammers building things like Walls and archers in a game that is highly offensive driven. An AI that would whip and draft as aggressively as a human would in the higher difficulties be unbeatable. In a system that rewards strategic maneuver as well as economy you get to actually fight the AI fairly and at a variety of different power levels rather than simply overcoming them with brute force.

Because, as Jon Shafer repeatedly said when asked about the early stages of the development process, "Civ4 is still a pretty good game." And it's hard to top. For the people who like the kind of gameplay that Civ4 has to offer, it would be very hard to create a successor that they like even more. Using Civ4:BTS as a base and building on top of it was also out of the question, since Civ4 was pretty confusing to many new players already, and adding more complexity on it would further limit the customer base. Hence, the only economically viable tactic is to do something that's still Civ, but different enough to tap into new target groups.

Civ4 is not a "failed experiment" and I see no reason why Firaxis should see a massively successful game as such. I think the opposite is true - Civ4 is/was too good to not pose a problem for its sequel. This isn't a new concept at all btw, although it's much better known in the movie/TV industry.

The thing is though Civ V is not merely different. It's an aggressive response to almost everything in Civ IV. Problem with Stacks of Doom? Remove stacking entirely. Religion and Spying too much extra bulk? The whole Shebang gets the chop. Gaming journalism is just an extension of marketing. He's not going to give an honest answer because Firaxis is still making money on their old titles.

Let me tell you that it was quite unusual. Horses in those days were mostly driven by nobility and upperclass. Horses didn't grow on trees, you know.
But never mind, i can see where this is going to. Against such logic, there is no reasoning.

Horses come in different degrees. A crappy pack horse is unsuited for combat but can still get you there. In the same way Grandmas smart car can get your behind to the grocery store and back, it gets you places faster than walking.

 

[Psyringe]

http://forums.somethingawful.com/showthread.php?threadid=3137775&userid=0&perpage=40&pagenumber=1

SlightlyMadman is the user name. Afraid you are going to have to control F through 116 pages but it's discussed IIRC.

Ahem. "Go find a needle in a haystack" is not the response I'm hoping for when I ask someone to put the evidence on the table. Oddly, googling for the username, the thread name, and "stack", doesn't lead to any of his posts: Check the search.

Also, judging from hs other Civ-related posts on the board, I do not get the impression that he is (or was) actually a developer at Firaxis. What's his real name?

That it's a rare event is in itself proof that the AI cannot handle stacks because it's not building enough of or the right kind of military composition.

Ah, okay - so in your eyes, the CIV4 AI fails because you say it has difficulties in a situation that rarely happens, and when called on that, you say that the situation rarely happens because the AI fails in creating it in the first place. This is quite a circular argument.

Currently I get the impression that whatever I say, your reaction will be "the Civ4 AI fails", padded with a makeshift argument that may even contradict previous points you made. This makes the discussion a nice intellectual exercise, but also utterly useless.

The thing is though Civ V is not merely different. It's an aggressive response to almost everything in Civ IV. Problem with Stacks of Doom? Remove stacking entirely. Religion and Spying too much extra bulk? The whole Shebang gets the chop. Gaming journalism is just an extension of marketing. He's not going to give an honest answer because Firaxis is still making money on their old titles.

If you present random remarks by a forum member at somethingawful, who you think to be a Firaxis developer, as striking evidence, and at the same time dismiss statements that the lead designer repeated several times (and that clearly contradict yours) as "marketing lies", then I suggest checking the criteria with which you evaluate arguments. The factor "Does this match my opinion?" may have greater impact than it should.

I'm not saying that everything that Shafer says is necessarily true, but I think you're dismissing this way too easily considering the dubious quality of other evidence which you obviously trust.

In my opinion, the "chopping" of Civ4 gameplay elements has two reasons. One, as they said, the necessity to produce something substantially different from Civ4. Two, the need to make the project manageable, considering that it's just the vanilla version, so some content from the complete version of the predecessor has to be cut. The changes to the rules system in (1) also made it more difficult to reuse old code, which exacerbated (2). Hence, the chopping of Civ4 gameplay is not a statement that Civ4 was "failed", but a necessary consequence from having to deviate from the previous formula due to Civ4 being hard to top for the people who liked it.

There's of course a lot of speculation in there, but it's at least in line with the official statements we do have, while you must assume that Firaxis feeds us lies to maintain your argument.

 

[Jediron]

Originally Posted by FatCatAttack
Horses come in different degrees. A crappy pack horse is unsuited for combat but can still get you there. In the same way Grandmas smart car can get your behind to the grocery store and back, it gets you places faster than walking.

You lived up to my expections, that's for sure. I would not have ask for more!
(iow: N>U>T>S>)

 

[Ex-Cop]

Excellent points OP. on the Mark

I suggest keep infinite stacking ala Civ 4. But add in the "flanking" Civ 5 bonus for adjacent units fighting. That way you can have SOD but also a bonus when another stack aides in flanking. Keep hexes.

 

[tom2050]

I honestly don't see what you mean. I look at Civ3 and see an AI that sends a flood of units across your territory one by one. I look at SMAC and see an AI that sends units against your cities piecemeal. I look at Civ4 and see an AI that assembles stacks, monitors their composition, moves them in a coordinated way, uses one type of unit to prepare the field for others (siege removing defenses), evaluates the attack chances and priorities of the units contained in them before committing to an attack order, and even builds stacks of naval transports (complete with escorts) to transport stacks of land units around.

So these are the things that make intelligent AI, huh?

Civ 3 AI does/has done all these things:
1. Assembles stacks.
2. Monitors their composition.
3. Moves them in a coordinated way.
4. Uses one type of unit to prepare the field for others.
5. Evaluates the attack chances of units
6. Builds stacks of naval ships.

doesn't mean it's smart AI.

 

[Psyringe]

So these are the things that make intelligent AI, huh?
(...)
doesn't mean it's smart AI.

Well, the things I listed are criteria which can be used to determine whether the AI can handle stacks (which was the question), wouldn't you agree? Obviously, an AI that can assemble stacks and monitor their composition (Civ4) handles them better as one that doesn't (SMAC). The same is true for the other criteria I listed.

Whether this constitutes a "smart" AI is a different question. Personally I'd say that creating an AI that can do these things decently is not trivial, and was quite a progress in the Civ series. But that wasn't the question here.

 

[ohioastronomy]

The thing is though Civ V is not merely different. It's an aggressive response to almost everything in Civ IV. Problem with Stacks of Doom? Remove stacking entirely. Religion and Spying too much extra bulk? The whole Shebang gets the chop.

You've just isolated the central problem with Civ 5 in a nutshell. Aggressive over-reaction in design leads to failure.

 

[Toady]

You've just isolated the central problem with Civ 5 in a nutshell. Aggressive over-reaction in design leads to failure.

The good ol' "sweep the corpse under the rug" tactic backfiring?! Imagine that!

 

[FatCatAttack]

Ahem. "Go find a needle in a haystack" is not the response I'm hoping for when I ask someone to put the evidence on the table. Oddly, googling for the username, the thread name, and "stack", doesn't lead to any of his posts: Check the search.

Also, judging from hs other Civ-related posts on the board, I do not get the impression that he is (or was) actually a developer at Firaxis. What's his real name?

No clue as to what his real name is. But the forum's mods have a history of confirming things like that and no one questioned him. Monster with 21 faces is the guy from the failed MMO APB for example. Others like the guy who worked on Left 4 Dead go by their real name.

Ah, okay - so in your eyes, the CIV4 AI fails because you say it has difficulties in a situation that rarely happens, and when called on that, you say that the situation rarely happens because the AI fails in creating it in the first place. This is quite a circular argument.

How is that circular? They are both problems with the AI and are directly related. A stack is essentially an expression of a Civ's economy as well as it's use in combat. The AI can't do either therefore it can't "handle stacks."

Currently I get the impression that whatever I say, your reaction will be "the Civ4 AI fails", padded with a makeshift argument that may even contradict previous points you made. This makes the discussion a nice intellectual exercise, but also utterly useless.

I don't see a contradiction. I do see that you seem to have more of a desire to score points than you do in anything else.

If you present random remarks by a forum member at somethingawful, who you think to be a Firaxis developer, as striking evidence, and at the same time dismiss statements that the lead designer repeated several times (and that clearly contradict yours) as "marketing lies", then I suggest checking the criteria with which you evaluate arguments. The factor "Does this match my opinion?" may have greater impact than it should.

I'm not saying that everything that Shafer says is necessarily true, but I think you're dismissing this way too easily considering the dubious quality of other evidence which you obviously trust.

In my opinion, the "chopping" of Civ4 gameplay elements has two reasons. One, as they said, the necessity to produce something substantially different from Civ4. Two, the need to make the project manageable, considering that it's just the vanilla version, so some content from the complete version of the predecessor has to be cut. The changes to the rules system in (1) also made it more difficult to reuse old code, which exacerbated (2). Hence, the chopping of Civ4 gameplay is not a statement that Civ4 was "failed", but a necessary consequence from having to deviate from the previous formula due to Civ4 being hard to top for the people who liked it.

There's of course a lot of speculation in there, but it's at least in line with the official statements we do have, while you must assume that Firaxis feeds us lies to maintain your argument.

Even if you don't trust the user the evidence points towards Civ IV not being seen in as favorable a light as you think it is. Generally when a sequel is seen as a success they don't fix what ain't broken. The Call of Duty series for instance provides the same basic cinematic shooter gameplay in each addition. They ADD to what's there in increments and they may do some rebalancing. Star Craft 2 is again a similar situation. Even keeping to the formula as much as they did some of the hardcore RTS pro-poopers still find sacrilege in what Blizzard has done. But they obviously felt that this audience was worth catering to or else they wouldn't have tried. Civ III (just from observation) is a clear build up from Civ II, but Civ IV doesn't have that relationship with 3 as 3 did with 2. Civ IV is a response to III.

Developers are going to lie. That's a given. I mean have they ever thrown Civ III under the bus? They say it's a "must have" for gamers right on their website

http://www.firaxis.com/games/game_detail.php?gameid=3

afterall.

Or Civ Rev? If we say hypothetically Civ V is a lost cause as the gloom and doomers claim will they say bad things about V? Brad Wardell's mea culpas for Elemental War of Magic made a splash in the gaming news for how unusual it was.

Bioware is going under a similar process. Their game Dragon Age was a big success. It was meant as a spiritual successor to Baldur's Gate but the sequel has more in common with action games like Dynasty Warriors. They have done some trash talking on the prequel though, although in a soft "please buy the game anyway" kind of manner. They saw the former game as a failure in design and threw large chunks of it away in response.

You've just isolated the central problem with Civ 5 in a nutshell. Aggressive over-reaction in design leads to failure.

Not always. Star Control 2 is VERY different from the original and is considered a gaming classic. Star Con 3 is relatively similar to 2 more than 2 is to one but is crapped on often. But I would say it is definitely a risky move and is prone to failure as you put it. That's why I figure they must not like Civ IV at all to make such radical changes.

 

[tsoky]

If you played Civ: Call to Power, the way they organized battles there was quite good. 9 units per tile max, and battle was all stack vs all stack.

 

[man-erg]

There's a dev who posts on the Somethingawful forums who wrote in a Megathread on Civ IV that they basically threw their hands up and accepted SoD's as an inevitability by BTS.

That's interesting, however I still don't see what is wrong with having *limited* stacking. Ideally based on terrain type. eg 2 units in hills, 4 in plains, 1 in desert. With hexes, surely this would allow for far more interesting and less frustrating combat that either SOD or 1upt? And it would be simple enough that the apparent target market of Civ 5 wouldn't have their brains frazzled. Add back in the "grouping all units on a tile" of Civ 4 and you have the best of all worlds - more believable combat, less micromanagement. Add a config option to toggle whether you wanted "1upt" or not...perfect!

Is it just that they want to keep it as rigid and simple as possible to make the AI coding easier?

 

[WuphonsReach]

Generally when a sequel is seen as a success they don't fix what ain't broken. The Call of Duty series for instance provides the same basic cinematic shooter gameplay in each addition. They ADD to what's there in increments and they may do some rebalancing.

That's funny. Did you ever play the original Medal of Honor: Allied Assault (MoHAA)? Did you play the original Call of Duty?

One of the big annoyances in MoHAA was infinitely spawning enemies until you managed to step on some invisible trigger point. When most of the MoHAA devs left to create the original Call of Duty, they specifically stated that one of their goals in level design was not to rely or ever use infinitely spawning enemies. And we ended up with some very good gameplay as a result.

Fast forward a few years - Call of Duty 2 comes out. Full of cheesy infinitely spawning enemies. Along with magical grenade indicators and being obviously aimed at console players. It was pretty obvious that those who had experience making good levels and good play were either no longer with the company or were not listened to.

(For the map makers, CoD:UO was pretty much the pinnacle of the series. Combined arms of tanks/infantry, maps that could be made larger then a shoebox, and games as large as 20v20. A lot of the newer shooters don't even let you have custom maps.)

 

[polypheus]

If you played Civ: Call to Power, the way they organized battles there was quite good. 9 units per tile max, and battle was all stack vs all stack.

I agree completely. Or to satisfy the people that want a Panzer General-like mini-game, they could have created a mini-map for battle resolution that allowed for some tactical control as well.

The thing about tactical control though is to recognize that the AI will always be at a disadvantage as it can never really coordinate well multiple units the way any decent human can. At least with SoD, it was able to fight on somewhat more equal footing as it is is simply easier to program it to assemble a stack and then move it as though it were one unit.

Its just so frustrating that Shafer and company did not research good proven solutions that have existed for decades and came up with this Panzer General 1UPT system on a strategic-level map. He had many to choose from that would have been much better than SoD and this 1UPT nonsense.

But back to the topic in particular: 1UPT is indeed a failed experiment. Even some of those that think highly of it due to the newness factor will quickly tire of the constant shuffling and micromanagement of repetitive formations and tactics that become rote and routine against an AI that simply can't a human's level of coordination.

 

[Lord Khorak]

I love the one unit per hex. The AI is dumb but I'm not going to bag a good mechanic just because in the very first interation the computers handling of it isn't brilliant. Might as well have stopped the Total War games a decade ago if that were the case, 'cause damn, the battle AI has always been an utter moron.

 

[lschnarch]

I love the one unit per hex. The AI is dumb but I'm not going to bag a good mechanic just because in the very first interation the computers handling of it isn't brilliant.

But that is exactly what many of the players complain about.

When your chosen game mechanics lead to the fact that the AI is weak, this decision is at the very least highly questionable.
A design decision should be checked against the abilities of the AI. If the AI cannot handle what the design requires then the decision was wrong.