Unforeseen Intersections: Half-Life As World Allegory

Begin a new game of Heroes of Might and Magic 3 (the best game in the series thus far). Play for a few turns, at least to the point that you have revealed the territory of your AI foes and may observe their movements unhindered by the fog of war. Save the game, end your turn, and note the movements of the AI. Then reload your game, end your turn, and observe once more. The AI will move along precisely the same paths and engage in precisely the same battles that it did previously. It's as though you've rewound a movie and watched the same scene a second time.

Compare this to certain scenes in Half-Life, in which the player may passively observe AI skirmishes between marines and alien forces. The marines may win on one play through, but reload the last save in order to watch the battle again and you will oftentimes find the results to be reversed. No two firefights are exactly alike, even when the same fight is reloaded again and again. The starting positions of the AI entities are the same; the level geometry is no different; the rules governing AI behavior do not change between reloads. So what is it that Half-Life has, and which Heroes 3 lacks?

Put simply, the rules that govern the behavior of Half-Life's AI are probabilistic, whereas those at work in Heroes 3 are deterministic. The distinction is simple, but its implications may help guide us toward a better kind of game -- and maybe even a better kind of worldview altogether.

Physical, or scientific, determinism (hereafter: determinism) is the doctrine which holds that, given knowledge of the position and momentum of every particle in a system at time X, as well as knowledge of the laws that govern the system, a sufficiently powerful mind may calculate the entire past and future behavior of the system as a whole. Determinism reached its popular apex in the eighteenth and nineteenth centuries, after Newton revolutionized the intellectual world with his physics and calculus. But ever since the Einsteinian revolution of the early twentieth century and the subsequent introduction of quantum mechanics, determinism has lost traction in the minds of physicists and philosophers, while yet maintaining some grasp upon the population at large. The movement has its proper roots in the philosophy of the ancient Greek atomist Democritus -- though even earlier mythological entities, such as the Homeric Fates, may serve as evidence that the problem of determinism has weighed heavily upon the Western mind since before the Dark Ages of ca. 1200-800 B.C.

If the determinists are right, and if every single fact in the universe is determined by physical law, then the future is every bit as fixed as the past, and the notion of free will must necessarily evaporate before our desperate eyes. (Although some philosophers believe that determinism is fully compatible with human free will; they are called compatibilists -- or, on occasion, crazy.) That which we call the present can have been produced by only one past, and shall yet produce only one future. Naturally, many philosophers over the years have found this apparent conflict between determinism and human agency most unsettling.

I cannot describe within a reasonable amount of space all of the many attempts that have been made to solve the problem of determinism throughout the ages; nor could I easily do justice to the many subtleties of the arguments, even if I were undaunted by their sheer number. Recently upon reading Charles Sanders Peirce's essay "The Doctrine of Necessity Examined" (P474: The Monist 2 (April 1892):321-27; also CP 6.35-65), however, I was struck by the following paragraph against determinism:

Peirce wrote:

2) By thus admitting pure spontaneity or life as a character of the universe, acting always and everywhere though restrained within narrow bounds by law, producing infinitesimal departures from law continually, and great ones with infinite infrequency, I account for all the variety and diversity in the universe, in the only sense in which the really sui generis and new can be said to be accounted for. The ordinary view has to admit the inexhaustible multitudinous variety of the world, has to admit that its mechanical law cannot account for this in the least, that variety can spring only from spontaneity, and yet denies without any evidence or reason the existence of this spontaneity, or else shoves it back to the beginning of time and supposes it dead ever since. The superior logic of my view appears to me not easily controverted.

I have learned the hard way time and again that Peirce was a really, really smart guy. So whenever I come upon a passage for which I have no frickin' idea what the man's going on about, I try to give him the benefit of the doubt, under the assumption that he's probably worked things out in a manner that has eluded me, and I am simply blind to his superior grasp of reason. Usually, after due time and much effort, I am able to discern the gist of his arguments; but sometimes this is not the case, and I must resign myself to a difference of opinion with my main man Charlie.

When I read the above passage, I initially felt certain that this would be one of those sad times of philosophical disjoint between myself and Peirce. For I could not fathom why Peirce thought that an indeterministic universe would necessarily allow for more variety in the world. Peirce seemed to be saying that a deterministic universe would be one without beauty -- or at least, without a certain important kind of beauty, i.e., that owing to nature's variety and spontaneity. But beautiful variety seemed to me to be accepted as inherent to both the deterministic and indeterministic views of the world; for determinism and indeterminism do not address different worlds as effect, but rather only the supposed contrary causes for the single world that all parties acknowledge we inhabit. It seemed that Peirce was not giving the determinist quite enough credit here.

Then I wondered: is it possible that the determinist and indeterminist do not merely disagree over the supposed cause of the world as we know it, as I initially had thought, but that they also disagree upon the very essence of that world itself? Perhaps they are articulating not merely conflicting world-rules, but also conflicting worlds themselves.

And that's when the analogy to games struck me like a bolt from the blue. For that is exactly the advantage that Half-Life has over Heroes 3: out of spontaneity, variety; out of variety, a kind of beauty. For is it not beautiful (in a sense of the word) to know that each outcome of battle is in some way randomly decided? Who will deny the surge of adrenaline that accompanies the full realization that anything might conceivably happen? Who is not intrigued by the fact that events are not pre-scripted according to some arcane algorithm, but are on the contrary loose and spontaneous? Who would not rather play a game wherein -- all other aspects being equal -- there is an aspect of indeterminacy to the AI?

I now see that Peirce's world is a world in which we may by default justify the delight we feel in studying the history of events, or in predicting their future course. I think that's pretty neat.

What's also neat is the fact that my experience as a computer gamer helped me to interpret this crucial bit of philosophy. Though it was released as long ago as 1998, Half-Life continues to shape me in ways that I cannot anticipate.

--Lobo

Comments

Random generators, like any other thing, are only good when used with good measure. For example, if, in a hypothetical game, you have a squad of 20 soldiers vs. 5 aliens of equal strength and AI, and there's a possible random scenario of the aliens defeating the soldiers - thats going way too far with the generators.
The losses may differ in the losses on the soldier's side but rest assured, all the aliens should be wiped out in any possible scenario.
Advance Wars battles, for instance, are a good example of a proper random generator. There's no chance in hell that a wounded tank will defeat two whole tanks, but there's a variation in the damage they take.

In short, the outcome should be still largely predictable based on general weight of the game balances, unless the sides are even to the point where it the outcome could be decided by the "margin of random".

A perfect example of a GOOD random generator use would be an RPG where NPCs roam on their own business. Say, a group of hunters is shooting wild rabbits. Each of them has a small arms skill that affects their aim with the rifle. One of them eventually misses and the projectile goes past the rabbit and into a gang member on the other side of the road.
As a result, the entire gang comes down on the hunter group and mows them down with machineguns.

If you look at just the beginning and the end of the encounter, it will seem like a random generator was used in a BAD way. Rabbits are still alive, hunters are all dead ! Impossible !

But if you look at the whole flow of events, it becomes clear that it was a large event caused by several small random influences, which altogether are not improbable. The hunters straying into gang territory was caused by semi-random movement of the rabbits, and the shootout was caused by the semi-random miss of one of the hunters.
In overall, this carefully measured random influence on the game factors has managed to create a truly dynamic, unscripted situation both the reasons for and outcome of which remains plausible.

I am not sure what my point is, but I always found this subject to be fascinating when it comes to theory of modeling of game worlds that seem to be "alive".

I personally think that a universe which does not allow for the beauty of chaos would suck.

That's my philosophical thought for the month.

The chaotic, random AI of Half Life has long been one of my favorite examples of game design done right. The moment you realize that the invisible ninjas are adapting to you is pure gaming bliss. We need more of those moments. We also need more views of the universe from inside the helmet of Lobo. Both are things that do not suck.

shihonage wrote:

Random generators, like any other thing, are only good when used with good measure. For example, if, in a hypothetical game, you have a squad of 20 soldiers vs. 5 aliens of equal strength and AI, and there's a possible random scenario of the aliens defeating the soldiers - thats going way too far with the generators.
The losses may differ in the losses on the soldier's side but rest assured, all the aliens should be wiped out in any possible scenario.

Excellent point. This is what Peirce meant when he wrote:

Peirce wrote:

By thus admitting pure spontaneity or life as a character of the universe, acting always and everywhere though restrained within narrow bounds by law, producing infinitesimal departures from law continually, and great ones with infinite infrequency...

In other words, in Peirce's view of the world, it's not like anything at all is allowed to happen. The rules that govern the world are fuzzy along the edges, but not nonexistent. I think this is the model to which games should aspire. Many gamers have come to value such a model in many other contexts than simply watching AI behavior, too. For example, flight simmers place great value upon the semi-random flow of a dynamic campaign.

Interesting article: anything which makes philosophy more relatable and digestible, in my mind, is a good thing. What better example to use than games? Now, I want to see what you think Nietzche would say about Megaman

Excellent article, Lobo! One small thing though:

Lobo wrote:

Put simply, the rules that govern the behavior of Half-Life's AI are probabilistic, whereas those at work in Heroes 3 are deterministic.

The rules for both are deterministic; Half-life's outcome, however, is probabilistic.

My own thoughts on the matter of AI are proving to be quite difficult to pen down coherently. Perhaps my impending bout of cold has something to do with it...

shihonage wrote:

...In overall, this carefully measured random influence on the game factors has managed to create a truly dynamic, unscripted situation both the reasons for and outcome of which remains plausible.

Exactly why I love this kind of thing. Going back to your tank scenario, while the crippled tank will most likely be destroyed, the varying amounts of damage it causes to the opposing tanks will influence future events, therefore opening up even more possibilities. It may have done enough to sway the next battle, or buy time for other units to make a tactical retreat.

Fletcher bridged the gap between sex and games, Lobo between philosophy ((in)determinism) and games. When oh when is that Nobel Prize coming GWJ's way?

But, chaos theory returns the universe back to (inescapable) determinism. If you were to not only save your half-life 2 game, but the exact condition of the universe with an infinite level of accuracy, and then play the same fight again, it would unfold EXACTLY like it did the first time you watched it. The only reason a non-deterministic view of the universe exists is because we lack the ability to model systems and measure them at the scale required to see the mechanics in play. Take that you notion of free will!

Have you read Asimov's Foundation series Lobo? Some interesting parallels between the underlying theme of the series and the thoughts you've laid out in this article. One of the best series of books I have ever read with an interesting take on determinism.

BadMojo wrote:

But, chaos theory returns the universe back to (inescapable) determinism. If you were to not only save your half-life 2 game, but the exact condition of the universe with an infinite level of accuracy, and then play the same fight again, it would unfold EXACTLY like it did the first time you watched it. The only reason a non-deterministic view of the universe exists is because we lack the ability to model systems and measure them at the scale required to see the mechanics in play. Take that you notion of free will! :)

Thats exactly the Newtonian notion that Quantum Mechanics puts a chink in. Spose you do turn back the whole universe when you load that game of Half Life but thanks to our friend, Einstein, this time an electron gets bumped a microsecond earlier than the last time. Half Life's random function gets seeded at 1821908781987 instead of 1821908781988 and a marine keeps firing as opposed to taking cover. The marine takes a shot to the head and now the aliens have the upper hand.

The best example of determinism and spontaneity being used together that I can think of is when Master Cheif takes out an Elite in Halo, the Grunts being commanded by that Elite always flee in fear (determinism) but do so randomly, er, in spontaneous directions.

It's like playing a Civ 3 game with random seed preserved versus one where it isn't.

Preserved: If you attack an enemy city with your stack of tanks and then reload the same saved game and try again with the exact same moves, the outcome will be the same.

Not preserved: The same series of moves may produce a more favorable outcome or a less favorable one.

My first encounter with Half-life's Marines remains a touchstone of sorts in my memories of past gaming experiences. I'll never forget the sense of surprise, followed by fear, followed by amazement, infused by a euphoric rush if adrenaline, when I heard the plink-plink of a grenade bouncing into my hiding place, tossed there by an apparently thinking marine who had the sense to take cover and flush me out.

Lobo wrote:

The rules that govern the world are fuzzy along the edges, but not nonexistent. I think this is the model to which games should aspire.

Exactly! I want to move through the world of a game whose rules can be learned, understood, and predicted, but not to the point that the elements of surprise and spontaneity are absent, or that the rules break down completely. The sight of an NPC stuck partway through a wall, running in place, is too great a departure from the rules. The NPC grabbing a grenade I threw and popping out of a bunker to chuck it right back at me, now that's exciting.

I think in terms of gaming, when we're talking about the rules, we're talking about the laws of gameplay, built into the code - whether deliberate gameplay contrivances or simply limitations of the medium - that we expect to govern our gaming experience. Half-life played with our expectations by providing experiences we wouldn't have predicted, both through unforeseen scripted events that appeared spontaneous, and enemy AI that was truly unpredictable.

Along those same lines, I think the wild-ass events that occasionally occur in Battlefield 2 are what make it such a terrific game. Experience with that game brings knowledge of the system of rules that govern cause and effect within the game environment - but no amount of familiarity with these rules can prepare you for the zaniness of actual player behavior, and the sublimely unexpected turns of events that inevitably occur.

Sooo... loading up a CS:S game with bots and watching them duke it out is the same as taking a class in Philosophy? Where do I apply for credit?

Great analysis, Lobo! The Beauty of the Unpredictable applied to a game that Did It Right (tm).

Personally I'm really looking forward as to how this approach to game design will play out in the next Elder Scrolls saga. If even half of what I've heard about the NPC's 'motivations' are true it will be a game to behold! I wonder... will the NPC's freak out if you stalk them as they go about their day to day routines? Slowly descending into the syrupy depths of madness as a grim figure stalks their every living moment? Bearing witness to their basest desires. Always watching... waiting... waiting for them to do 'something!' Gasping as they wander into shops to find sustenance, muttering in awe as they go about their daily lives to scrape together enough gold for their next meal. Giggling with delight as they scurry to take shelter when an errant fireball nips at their heels. Haughty dismissal turning to horrified dismay as their scarred shadow creeps closer and closer gazing solemnly at their faces. Reading the hidden meanings and urges there. (And then gutting them in some black alley in the hopes that they were carrying an enchanted item)

Rezzy wrote:

I wonder... will the NPC's freak out if you stalk them as they go about their day to day routines? Slowly descending into the syrupy depths of madness as a grim figure stalks their every living moment? Bearing witness to their basest desires.

Important safety tip: Don't invite Rezzy to the next GWJ Slap & Tickle.

Copingsaw wrote:

Have you read Asimov's Foundation series Lobo? Some interesting parallels between the underlying theme of the series and the thoughts you've laid out in this article. One of the best series of books I have ever read with an interesting take on determinism.

I read the first three books in the series way back in middle school. I don't remember much discussion about determinism, but I was pretty young at the time so it may not have stuck with me. I've been planning to read/reread the entire series anyway, though. Thanks for the recommendation.

danjoolivaw wrote:

Thats exactly the Newtonian notion that Quantum Mechanics puts a chink in. Spose you do turn back the whole universe when you load that game of Half Life but thanks to our friend, Einstein, this time an electron gets bumped a microsecond earlier than the last time. Half Life's random function gets seeded at 1821908781987 instead of 1821908781988 and a marine keeps firing as opposed to taking cover. The marine takes a shot to the head and now the aliens have the upper hand.

You're just not measuring things accurately enough if you're not reproducing the same interaction from your save. I might not trust you to backup my universe.

Lobo wrote:
Copingsaw wrote:

Have you read Asimov's Foundation series Lobo? Some interesting parallels between the underlying theme of the series and the thoughts you've laid out in this article. One of the best series of books I have ever read with an interesting take on determinism.

I read the first three books in the series way back in middle school. I don't remember much discussion about determinism, but I was pretty young at the time so it may not have stuck with me. I've been planning to read/reread the entire series anyway, though. Thanks for the recommendation.

The first three are the good ones. There's a couple more written by Asimov after those three which aren't bad but sort of deviate and really start to get bad toward the end. Then there's some more written by people who aren't Asimov, don't read those.

Rezzy wrote:

creepy story

The real question is, will they notice if you're not wearing pants?

BadMojo wrote:

You're just not measuring things accurately enough if you're not reproducing the same interaction from your save.

One of the problems with the basic determinist position is that it's impossible to verify conclusively. To do so would require that we be capable of making measurements of infinite precision, and of calculating their implications with complete exactitude. There's no reason to believe that science can ever aspire to such goals. (And if the quantum mechanicists are right -- and the course of science since the early 20th century suggests that they are -- then determinism wouldn't hold even then!) To adhere to determinism would therefore seem to require some element of question-begging.

But the same could be said about indeterminism as well; for how could one ever know with certainty that a given event could not have been predicted using prior data, if only sufficient data had been available? However, the pragmatic approach allows for a rational shortcut to be taken around this difficulty. The pragmatist would argue that since, practically speaking, we can only ever estimate both the state of a system at any one time, as well as the natural laws that govern said system, then if there is some solid grounding for knowledge that lies forever just outside our grasp (as the determinist contends), then that grounding may as well not exist at all, since our experience will never lead us to it. Under a pragmatic conception of reality, that which by definition fails to ever enter into the course of human experience simply does not exist. There is no such thing as a necessarily and eternally hidden truth.

Lobo wrote:

But the same could be said about indeterminism as well; for how could one ever know with certainty that a given event could not have been predicted using prior data, if only sufficient data had been available? However, the pragmatic approach allows for a rational shortcut to be taken around this difficulty. The pragmatist would argue that since, practically speaking, we can only ever estimate both the state of a system at any one time, as well as the natural laws that govern said system, then if there is some solid grounding for knowledge that lies forever just outside our grasp (as the determinist contends), then that grounding may as well not exist at all, since our experience will never lead us to it. Under a pragmatic conception of reality, that which by definition fails to ever enter into the course of human experience simply does not exist. There is no such thing as a necessarily and eternally hidden truth.

Truly, you have a dizzying intellect.

I don't agree that just because you can't experience it means it doesn't exist or isn't true.

I don't agree that just because you can't experience it means it doesn't exist or isn't true.

That's not quite what he said.

The way I see it, you can argue the existance of either viewpoint ... mostly because they both are real. While I might know that the universe is more deterministic as measurability approaches infinity, I realize that the end of said function is not going to be reached. I don't see a conflict in the two aspects of how things work. Reconciling a perceived ultimate truth with the reality of the world is part of life. And fun at that.

this is without any doubt the most interesting thread I've read here until now. Let me bow humbly before He that is Lobo.

I don't think the analogy fits the differences between these two games. HOMM3 is simply exhibiting what many computer science majors find out their first year: It is notoriously hard to get a computer to generate a truly random number, and while we see it all the time, in a music playlist for example, it's because the problem has been solved by "some arcane algorithm" that people are reusing. Since strategy games are by their very nature "less" random, I would not be surprised if they repeats strategies, ie, go for the same high priority target each time. What I don't like about games like HOMM and Civ is a) they usually don't have any kind of grand strategy, all the units move with a mob mentality and no central intelligence, and b) they don't learn your habits and use that to outwit you.

Also, remember that the Half-Life marine has a relatively simple life to live. If he sees you, he's gonna come after you, one way or another. He's not going to attempt to trade with you, or start diplomacy, or spy on you, or run back up the stairs and build a base to pump out unicorns. He's not going to see you then go shopping, or turn and play patty-cake with his buddy. He lives in a 3D environment that is known and not randomly generated.

Perhaps most relevant to this discussion is the fact that computers themselves are totally deterministic machines, while humans live probabilistic existences, at least from the layman's point of view. "There is no computer error, only human error." Computers are designed to eliminate any sense of randomness, to perform the same strategy again and again. Humans will most like never do exactly the same thing twice.

So the idea of a computer generated randon number is itself a contradiction. How most computers solve it is usually based on the most random element in its environment - it's human user. It looks to the system clock (when you start using it) or some busy spot of memory (what you're using it for). I wonder if, in a controlled environment, a computer could play itself in Half-Life and everything would remain the same, game after game.

I believe good AI merely suspends your disbelief.

Interesting post, souldaddy. I'm curious as to what makes you think that in order to have indeterminacy in a computer game, a truly random number generator is necessary. All that is needed to fulfill the condition of indeterminacy is for more than one outcome to be possible given certain initial conditions. The outcomes themselves need not have an equal probability attached to each. We may even be able to predict the eventual outcome with a better than 50% chance of success; but so long as we do not arrive at 100%, the game will remain indeterminate.

souldaddy wrote:

I wonder if, in a controlled environment, a computer could play itself in Half-Life and everything would remain the same, game after game.

That's not what happens when the player observes an AI battle from afar, without participating in it. Does this fulfill your conditions for a controlled environment?

I've come into this discussion quite late, but Bungie's Myth 1 & 2 should be mentioned here.

The Myth games were Real-time tactical games with an extremely robust physics system, leading to a gameplay dynamic that felt "random" and "chaotic," like a good tactical shooter. In fact, the Myth games remain my gold standard for that sort of random, non-deterministic gameplay.

The games themselves though were entirely, 100%, deterministic. The save games seemed to exist solely as a record of the precise player inputs to the simulation. When reloading a game, the application would rerun the simulation at high speed, applying the player inputs to the simulation, secure that the end result of this process would put the player back exactly to the point where the game had been saved.

What happened from that point on, once the game had been reloaded, SEEMED totally chaotic. The game would never play out into the future the same way twice. The reason for this was that the granularity of the simulation was so fine that human reflexes could never hope to reproduce the exact same inputs two times in a row.

In essence, indeterminism is an illusion, but that's okay because to the observer the granularity of the simulation that is Reality provides a convincing sence of indeterminism.

polq37 wrote:

What happened from that point on, once the game had been reloaded, SEEMED totally chaotic. The game would never play out into the future the same way twice. The reason for this was that the granularity of the simulation was so fine that human reflexes could never hope to reproduce the exact same inputs two times in a row.

This thought should of course bring to mind Half-Life 2, and really any game with physics complicated enough to make recreation essentially impossible. Remember the part in the sewers where you wanted to take out headchopper thingys by blowing up a bunch of explosive barrels? I tried for quite some time to do it "The Right Way" but simply couldn't get any sort of consistency - move half a degree to left when shooting, and a completely different explosion occured. Great stuff.

Lobo wrote:

All that is needed to fulfill the condition of indeterminacy is for more than one outcome to be possible given certain initial conditions. The outcomes themselves need not have an equal probability attached to each. We may even be able to predict the eventual outcome with a better than 50% chance of success; but so long as we do not arrive at 100%, the game will remain indeterminate.

Yes, but how is the computer to choose which outcome to pursue? What basis should it make its decision on? Since a computer boils any experience you see down to numbers, if it calcutates Choice A as having a 60% chance for success, while Choice B having 40%, it's very easy to see what decision it will make. Thus, in HOMM, it always attacks the same city, thus becomes predictible. Even if the outcome of the attack is always different, the computer's choices are not.

I've been busting my head to think of some example in which Man has made a tool to do "A" and then turned right around and used that tool do the opposite of "A". Maybe parachuting, or race-drifting; modern cars are designed in everyway possible to prevent the rear end from cutting loose, but there is a sport to do just that. Computer AI is in many way just such an example. We strive for the illusion of a thinking, probablistic being behind that screen, but in reality it is a mindless, utterly deterministic machine. It's all about the illusion, about suspending disbelief.

To make up for that fact, we have gone to making more and more complex simulations, and thus hope to disguise our simple opponent behind the chaos in the game. This takes us back to your Half-Life battle. The starting positions of those monsters are the same every time, their AI algorithms are also the same, so the only real difference are the choices the computer makes at random, maybe what target to attack, or did this bullet hit or miss. It's a very little thing, but in such a big battle it means that the dominos fall in a different order each time.

The complexity of a modern PC is the reason you saw different outcomes as a passive observer. In fact, you were interacting with the battle in ways you weren't even aware of. In Staats' example, the game measured his shot in finer detail than his hand and his mouse were able to reproduce, despite his efforts and intensions. In your saved game, you probably loaded it at a different time of the day (even seconds are huge when a computer thinks in milliseconds), or maybe you had iTunes running in the background the 2nd time when the first you had BitTorrent. This means the system clock and the memory register where different each time the battle played out. Those are two common locations a computer pulls a seed for a random number generator. If the computer controlled its own environment, started a game at exactly the same time with exactly the same data loaded in memory, I think the battle would be the same every time.

Another interesting post. I hope you're planning to become a regular here!

Here's a question: if such minor factors as the state of the system clock or running background processes can have an effect on a game like Half-Life, thereby changing the way the game behaves, then why do they seem to have no effect at all on a game like Heroes 3? Isn't this an indication that Heroes 3 is more deterministic by design than Half-Life?

Lobo wrote:

Here's a question: if such minor factors as the state of the system clock or running background processes can have an effect on a game like Half-Life, thereby changing the way the game behaves, then why do they seem to have no effect at all on a game like Heroes 3? Isn't this an indication that Heroes 3 is more deterministic by design than Half-Life?

I would say that because Heroes 3 is strategy (and top-down, at that) it's going to offer clearer "paths of value". Coupled with a gamewold that's much more static than HL2's, as well as being turn-based, you're going to get a higher degree of consistency when you force the CPU to remake its decisions.

I hope that makes sense.

I agree with your analysis, Lord_Xan. In fact, that's precisely the point toward which I was gesturing.