IA Logo

IA Information

Dave Mark's books

Post-Play'em - Observations on Game AI

Dragon Age: Origins – Don’t Crowd!

May 25th, 2010

OK, so it has been a little longer than “later this week” for getting to Dragon Age: Origins commentary. I can’t believe I wrote my first look post back in late February. Of course what that really means is that I’m burning some serious GameFly time on this one title… and it’s not like I’m playing it much. My son has managed to get all the way through it, however.

Still, there are some things I noticed as I was going through (which, admittedly, is not all that far yet). One was an odd happening in the Ostagar camp. I was cruising around looking at things and talking to people. As I was standing there talking to Duncan (and a couple of others), a camp soldier of some sort came and stood near us. From the looks of it, he was supposed to be dealing with another person there, if I recall correctly. This didn’t really catch my notice until a 2nd soldier came up and attempted to stand in the same spot and talk to the same person. They jostled each other for a bit until the first one walked away – apparently on a random patrol route.

…a 2nd soldier came up and attempted to stand in the same spot…

I believe that it was the random patrol routes that caused this issue in the first place. My guess is that the guard logic simply selects a new destination and heads out in that direction. If, for some reason, there is someone already there, it doesn’t detect this or care… it simply continues to try to get to the node.

There are a few ways to either avoid that or fix it. First, the goal nodes could be reserved and released as they are selected and then left. The problem with this solution is that a goal node would be “out of commission” from the moment someone reserved it to the time they left it. If they were coming from all the way across the map, that node would be off the market for quite some time without actually being used. Additionally, it would prevent someone from heading in that direction until the original occupant actually left.

A lot of work to make the camp “feel alive” went down the drain…

A more realistic solution would be to simply be aware of other people. If there is someone on your “spot”, don’t actually try to move onto it. This could be done by simply stopping short of your intended goal (e.g. stand next to him) or by using some local avoidance. The later is probably better because collision avoidance should be part of the movement scheme anyway for other reasons.

Anyway, the result of this was that it was a jarring cancellation of my suspension of disbelief. Because it was right next to me and in view, I couldn’t help but notice that these two dolts were climbing all over each other. A lot of work to make the camp “feel alive” went down the drain right there.

Migration Completed (finally!)

May 24th, 2010

Well, after being dead in the water after The Great Blogger Screw-over of 2010, I’m now on a new web host and using WordPress instead.

I’ve got a backlog of games that I have been playing and making observations on — some with the help of my son, Aaron, who has caught the spirit of noticing AI issues. I’m hoping to put these all up on here in the next week or so.

Anyway, we’re back. Make sure you subscribe to Post-Play’em in your manner of choice. (As of this writing, I still haven’t put any “subscribe” link on here.)

Migration from Blogger to WordPress

May 1st, 2010

Well, after the migration to the “New Blogger”, I discovered that they won’t support my php includes so that the blog can fit into the rest of my site. Therefore, I am going to ditch Blogger entirely.

After having worked as an author for a couple of other blogs that used WordPress, I figure that it was a decent solution. Therefore, I am going to begin installing WordPress and see if I can get all my existing content over to it. Things may look ugly for a bit but hopefully I can get it taken care of.
Oh… and a big thank you to Google/Blogger for screwing up dozens (if not scores) of hours of my work. Good work, folks.

Blog Migration

April 30th, 2010

Because Blogger is doing away with their FTP publishing as of May 1st, I have to convert this blog over to a new type of system with them. Unfortunately, because I have their custom code embedded in my Intrinsic Algorithm pages, this is likely not going to go smoothly. Hopefully, there will be very little down-time. Please accept our apologies if something goes amiss.

Dragon Age: Origins – First Look

February 22nd, 2010

Just sat down with Dragon Age: Origin for a while. I know I won’t get too far into it anytime soon because, for a Bioware game, that usually means a commitment of a few months. I plan on blasting through more soon, but the 360 is on the same 43″ HDTV that the wife wants to watch Olympic stuff on. (I can’t begrudge her that since she usually doesn’t watch enough TV to get in the way of my gaming.)

To give a frame of reference of how far I am, I started as a City Elf and am only just about ready to roll out of town with Duncan. That has, at least, given me plenty of opportunity to go through both peaceful and combative encounters. Also, I haven’t yet dealt with the NPC behavior customization. I imagine that will be more important as I go foward, but for now I just wanted to live by the defaults.
Well Lookie Here!
One of the first things I noticed was when I was in a conversation. While I like fact that people’s eyes move and blink, I would like to see a little more movement. There is so much that can be said by breaking of eye contact, turning the head while still speaking, etc. I have written about this here and there before (and really should write a formal article about it). The point being, what they have done is a start, but it isn’t enough to really sell the character. As an example of dialog scenes that are well done, I point you to Assassin’s Creed (the original at least… haven’t seen AC2 yet). Previously, I wrote about how I liked that, in Assassin’s Creed, not only could you walk around during conversations, but that the NPCs did as well. They looked down, away, etc. I would like to see more of this. The walking around isn’t even quite as important as the head and eye control. Some of those can even be randomized — especially with parameters — so it’s not like it would take a lot of work or resources.

On the subject of head movement, however, one thing I noticed the first time that I was in a conversation with two people was that their head turns are NOT animated by hand but are automated. The reason I know this is that there is a very distinct start and end to the turn and that the turn rate is entirely constant during the whole turn. Put another way, if someone turned from looking at me to the person standing beside them, that 90° turn would go from a rotation rate of 0 straight to n, rotate that 90°, then change rate from n to 0 again. The rate, n, was always the same throughout one turn and on every turn. The overall effect made it look like one of those audio-animatronic characters in a Disneyworld exhibit. It was very mechanical-looking.
While I like the idea behind the technique, I believe a better way would have been to randomize the rate, a rate of change into the turn, and a rate of change out of the turn. This not only changes things up between one turn and the next, but also smooths out the starts and stops so it doesn’t look like little servo motors kicking in.
… one of those audio-animatronic characters in a Disneyworld exhibit.
Also, you don’t always have to look directly at someone. If someone standing next to you is speaking, you can turn your head part of the way and do the rest with your eyes. In fact, it is very common for humans to “split the difference” and look somewhere between the two speakers in that arrangement. Toss in the occasional look at other objects of interest in the environment (especially if you’re nervous, wary, etc.) and you can really sell a conversation without a lot of extra work.
Don’t Mind Me
She never acknowledged that I was even there.
While I appreciate that not every denizen of a place should attack you, I would at least like them to acknowledge my existence. For example, I was a little startled when I was storming the castle. In a couple of the rooms, I came upon a woman scrubbing the floor. She didn’t look up when I entered (if she was a servant, shouldn’t she care to see if it is a boss?), she didn’t look up at I moved around, she didn’t look up as I stood with my boots under her nose. She never acknowledged that I was even there. She didn’t do anything other than scrub the floor. It was a little startling, especially when everyone else in the place (friend or foe) recognized that, as an armed elf, I was seriously out of place. (Incidentally, I got a kick out of the fact that this matched up nicely with the theme of AIGameDev‘s AI Marmelade – their twist on the Global Game Jam. Their coverage of it was sub-titled “on little old cleaning ladies in stealth games.”)
I admit that this is likely a content issue. Developers would love to add more background characters to an area, but then you run into the issue of having them react in meaningful ways. Also, if those reactions are going to involve custom animation or sound, the expense goes up significantly. It would be nice to have a general list of generalized actions and interactions, however. Put people in there moving from place to place, carrying things, tending things, etc. When I come in, at least have them look at me. If I try to engage them, make them politely dismiss me. If I do something flashy like attack people, make them run. That much, at least, puts a little life into the place. Oh well… no one can ever say Bioware skimps on content. They just put it into length of the experience rather than the depth of it, I suppose.
Combat AI – So Far…
I was a little wary going into this first play session because someone I trust on AI commented that he was frustrated with his party members. As I said at the top, I haven’t gotten too deeply into the game and haven’t messed with the customization yet. However, there was nothing that I saw yet that was really “pants-on-head retarded.” I was a fighter as was my buddy. He attacked when I did or when we were attacked. I was too busy trying to do my own gig to see if he used a healing option when he should have. I don’t think he did because he “died” a couple of times during our great castle adventure. I didn’t think to give him a ranged weapon to see if he would switch at appropriate times, though. And now, of course, he is no longer in my party (taking some cool gear into the abyss of retirement with him). My next play session will have me leaving town with Duncan and likely teaming up with some dudes. We’ll see how things go then.
… it looked a lot like a tool that an AI programmer would make for a designer.
I only briefly looked at the customization system. From what I could tell, it looked interesting if you are an AI programmer or a designer. In fact, it looked a lot like a tool that an AI programmer would make for a designer. I would lay money that it’s origin is not too far from that. I can see how it would tend to quell the inevitable gamer uprising that would like more control of the autonomy of their party-members (is there a contradiction in there somewhere?). However, it was said by a few people that you actually have to tweak this system by hand to get people to work together properly. I really don’t like this idea. In my opinion, if it doesn’t work intelligently with the default settings, there is something wrong.
From a technical standpoint, the system looks like it is a parameterized rule-based system of some sort. I’ll have to play with it… for all I know, it is tweaking a behavior tree behind the scenes. Much of my conclusion on what is going on would be based on whether you can sort your rules in a particular order of consideration. Like I said, I only just glanced at it. More on this after I play. In the mean time, I refuse to look up any technical documentation on it just to see if I can figure it out on my own.
The enemies were pretty much standard “fight to the death” fare so far. They were all castle guards and punks, though. That would be right up their alley. I’ll have to see what other types of enemies offer later on.
Anyway, despite their similarities in production, the game is much more enjoyable to me than Mass Effect was — at least for now. I’d be pleased if that stayed the case.
Definitely keep an eye on the rss feed here. I’ll be writing more about this game later in the week.

Brütal Legend’s RTS elements

February 16th, 2010

Ok… now I get what Tara Teich is going to be talking about in her AI Summit lecture on Brütal Legend. She is going to be focusing on the challenges that they faced when they introduced RTS elements into a brawler-style game. When I did my first look, I hadn’t really seen evidence of how the RTS/brawler cross-over was happening. I just now played the first major battle (after you set up the stage for the first time), and I really see the mechanism at work.

Trying to manage all the different types of “units” you can summon makes for an interesting gameplay element. What’s more, as a player, you have to switch between solo brawling and teaming up with the other units. For example, you can get together with a group of headbangers and form a moving moshpit. At times, however, you have to make the decision to disband the group and go back to mashing solo, or team up with another group. Oh yeah, and you can still just issue general orders to units you aren’t teamed up with — orders like “come to me”, “stay and defend” or “attack that”. This can be done in addition to the local control as a team.
If it is difficult for me to manage what is supposed to be happening in the grand melee of units, I can imagine that trying to organize it properly from an AI standpoint must have been challenging as well. Because the units need to be nearby in order to be ordered or paired up with, I wonder if the units try to stay close to the player?

But what happens if there are legitimate people to kill that I’m not near? I can see frustration developing if the units were to just go off and attack on their own. Sure, it’s the right thing to do from a tactical standpoint, but the whole point is to give the player the sense of command. In a way it combines the autonomy of RTS units with a little bit of companion AI from the likes of Left 4 Dead or Mass Effect.
The other question is if they are doing combined tactics when they are on their own. I don’t mean advanced stuff like flanking, etc. Rather, are they just picking the nearest enemy or are they spreading out for maximum coverage?
On both of these points, I will have to do some more observation of what’s going on — and that usually means watching one of my kids play instead of doing it myself. They usually don’t argue much when I tell them it’s their turn.

Brütal Legend: First Look

February 14th, 2010

Next on the pre-GDC flush of my GameFly queue is the metal-themed game, Brütal Legend from Double Fine. This one has a little more urgency in it because my friend and AI Game Programmers Guild colleague, Tara Teich, did a lot of the AI for it. She is going to be explaining some of the challenges she faced in her lecture at the 2010 AI Summit at GDC. Specifically, she is going to be talking about marrying some facets of an RTS game (always a little more difficult on a console) with those of a typical brawler game.

Anyway, from a purely thematic standpoint, I have to find myself laughing rather often. It must have taken quite some time for the design team to make a list of every single hard rock and metal sterotype… or should I say stërëötÿpë… and figure out a way to insert it into the game. Anyway… that’s design, not AI, but it was worth noting. Very amusing game so far.

Anyway, from an AI-focussed view, it is certainly in a different ballpark than some of the other games I have been reviewing on here. RTS games like Company of Heroes, RPGs with heavy AI components such as Fable 2, and animation masterpieces like Assassin’s Creed are on a completely different wavelength than something like Brütal Legend. It’s not meant to be complicated, intense or clever. Instead, it’s meant to be accessible.

Give it Away, Give it Away, Give it Away Now…

One thing that I noticed just before I sat down here at the computer was the “tells” during a “boss battle” with a giant metal spider. As is typical in boss battles, the enemy will have a certain list of attacks that it performs on you. Each one of those attacks is preceded by a short preparatory move so that you know it is coming. There’s not a lot of subtlety or guile in boss battle “tells”. That’s the whole point. The other aspect of having tells is that you can pretty much string together attacks in any random order because the only thing the player is looking for is tell -> attack. On the other hand, I thought I detected a distinct pattern of some of the attacks. For those of you who have fought that spider, it seemed like there was a sequence of
  1. release the little spiders
  2. shoot the splashy stuff
  3. the engine gets exposed
It happened enough times that I could eventually put together the sequence of dodging the splashy stuff, tossing off the little spiders, and then shocking the engine.
To be honest, though, I really don’t like this type of combat. I’m not a puzzle-solving kind of gamer. I’m more of an NP-rediculous, react, plan, solve, be creative with how to ass kick, kind of guy. However, that’s a legit design decision for the majority of the gaming public and I’m cool with that.
All Together…

I can see where Tara is going with her idea of merging RTS and brawler features, though. When you start fighting along with the headbangers (some of whom I swear I went to college with), the game introduces you to a few rudimentary commands such as follow me, stay here, attack that, etc. Those aren’t so intriguing from an AI standpoint. A little further on, however, you are introduced to the idea of the “mosh pit” – another wonderful twist on a metal cliché. When you are surrounded by 6 of the headbangers and they form the moving circle around you, it is amusing. When you actually enter combat with this ring of cranium-wielding warriors, it becomes interesting.
I wish I had video of these battles so I could slow it down. I want to see how much intelligence is going on with their respective target selection. It seems like they are doing more than just hitting anything that comes near them. They do go out and find a near-by target, I think. However, do they divvy those targets up or is there occasionally some inefficient overlap where 2 ‘bangers go after the same guy? Dunno… This would have to be done with some sort of sorting and semaphore system — in layman’s terms, each one saying “I got this guy here… find your own”. I simply don’t know if that would be worth going through the trouble, however.

The animation during the cutscenes is nice. The faces (especially Jack/Eddy) are expressive in an amusing way. I have to assume that is hand-done, however. Not much of an AI-fest there.
The rest of the animation is rather typical. Again, they aren’t trying to be Assassin’s Creed here. I’m OK with that.
I think what I’m looking forward to most as I move through the game is more of the group scenes. I want to see how the groups of headbangers (or whomever) deal with larger groups of the enemies. After all, that’s what Tara identified as a challenge of sorts.
In the mean time, I’m just gonna rock out!

Odd Glitch in Company of Heroes

February 11th, 2010

Just to be clear, I really like Company of Heroes. I have always been an RTS fan since the original Warcraft and progressing through Starcraft, Age of Empires/Kings and especially Empire Earth. That being said, CoH is very well done.

I only recently got around to grabbing the full version of the original CoH and the expansion, Opposing Fronts. (Thank you Valve weekend sales!) My prior posts here were only about the downloadable demo. Obviously, there are some significant differences in the full release version. I like them all.
Rather than continuing through the script-fest that campaign maps usually are, I have been playing a lot of skirmishes against the computer. I find a lot of the tactics by the computer to be pretty familiar for RTS games.
For example, as I mentioned in a prior post, it is obvious that there is a lot of influence mapping going on. If I find myself a little behind on producing units, I will often be drawn into a series of dashing from defense point to defense point as the computer picks on whatever is currently the weak spot. My only question in this situation is how much is the AI cheating with knowledge about my force disposition. I know that it is not that easy to discern what the computer has placed and entrenched… it seems like he knows where my weak spots are even if they are deep within my territory.
Anyway, as I was playing a skirmish map, I ran across something very odd. Normally, I had been playing against the “normal” level of difficulty. I had been winning fairly regularly against that level, however, and decided to ratchet it up one level. I was playing the Lyon map against the computer on “hard.”
Because this was a control point map, I was quite used to meeting the “normal” AI at the river as we battled for those 3 points. I figured that the “hard” AI would be just as quick to get to the river. However, I captured the first point… then the second… without seeing the AI at all. I began to suspect that either something was dreadfully wrong or that “hard” used a different tactic than “normal” – such as “build up and crush him” rather than “dash for the control points”. Not too long after that, it became obvious that something was indeed seriously amiss.
Thanks to the wonders of replay files, I was able to go back and witness what the computer did… or did not do. A video is included below.
I sent the video privately to my friend and colleague, Chris Jurney, who did much of the excellent AI work on CoH. He thought it was amusing as well. His speculation was that the Lua script exploded somehow. It obviously didn’t cease working entirely because we see some orders trickling through at times. He invited me to send him the replay file so that he could forward it on to the Relic programmer who is responsible for it now.
It doesn’t seem to be a problem with the map itself as I did go on to play a new match on that map on “hard” and it proceeded to hand me my buttocks as I expected. On the other hand, it may be a problem with a combination of that map, that difficulty level, and a randomly selected tactical approach that knocked a variable out of bounds or something. Regardless of the cause, this is a great example of how important testing is… and yet things still slip through.
Anyway, I will try to do more video captures of some games later – especially the view that shows how the computer is moving throughout the map during the game. That has been very interesting to analyze because it shows how the influence mapping and even the pathfinding is working as part of the decision making.

Mass Effect: First Look

January 29th, 2010

I’ve had my copy of Mass Effect (yes, the original) from GameFly for a month or so now. My son played it for a few straight weekends when he was visiting and finally finished it. Seeing how long it took him to go through has convinced me that I do not have the time to ever complete the game. However, I did want to get a feel for what it was like specifically for reference purposes. Here’s a few of my first impressions.

Good Dialog Animation

One thing I noticed that was rather pleasant was that the cut scene dialog animation was reasonably well done. This is even more impressive since there are at least a dozen races that I have encountered so far and all of them look believable when they are speaking. I happen to have a soft spot for ol’ Urdnot Wrex who looks particularly compelling when he speaks. Kudos to BioWare for pulling that off well.

Fixed Cut Scenes
On a down side of that is the fact that I now feel really rooted to the ground during those dialogs. Much of this is by an unfair comparison, however. After playing Assassin’s Creed where I could wander around aimlessly during dialogs, I found myself a little restrained during the Mass Effect ones (of which there are plenty). Which brings me to another point…

Odd Disconnects
Having been a veteran of Neverwinter Nights, I am well familiar with the extent that BioWare can take their contextual dialog trees. While I agree that this is still largely a necessary evil in RPG games, there are times when I am a little alarmed by a continuity disconnect.
For example, upon arriving in Noveria, I had a pleasant chat with the lady Gianna. She tells me upstairs at the security station to come see her to have an appointment with Anoleis. It took me less than a minute to get down to her office where she greeted me with the question, “What do you want?” It was rather startling to have just left her and have her forget that she told me to come ask her something specific. Of course, this is something that can be chalked up in part to poor writing. On the other hand, this is also something that could have been added as a trigger flag in the dialog tree itself. That is, if the player gets to you quickly, acknowledge it and say something appropriate. In this case, it would have been “Ah, you are here. Let me announce you to Anoleis.”
On the other hand, the in-dialog flags seem to be well done. Often, they will remember that I have asked about one topic in another branch of the tree entirely. That is well done. Of course, I was tickled when one character told me off by saying we have already discussed something. On the other hand, I can see how that would be annoying if you were distracted by something and didn’t hear what was said.
Another peculiar interaction was when Gianna met me outside a room after I had finished a combat sequence. (I believe it was outside that one dude’s office.) She told me to talk to her in the hotel bar before I spoke to the guy who sent me on the mission. Uh… OK. The problem was, as soon as the cutscene was over, she was gone. We didn’t get to see her walking away or going to the bar, etc. Not having anything else to do, I went to the bar and found her already there (she’s quick!). Again, this effect is a by-product of having the cutscenes be somewhat de-coupled from the actual live action gameplay. It’s mildly annoying.
Not much of that has to do with AI, however. This next bit does.
Combat AI
In a few of the combats that I have been in, I’ve noticed a peculiar combination of tactics. While I praise the design that puts some enemies behind cover for a while. It seems like they also leave cover for no reason, walk some pre-programmed path while still firing at me, then return to cover. If that doesn’t smack of a shooting gallery, I don’t know what does. Additionally, there have been times when an enemy inexplicably leaves cover and runs right up to me in the open. It seems to be a design decision to “mix things up” however. While mixing things up is fine, it doesn’t really make a lot of sense from a combat standpoint. Because of that, it makes the enemy look dumb. If the enemy looks dumb, people think the AI is dumb. It is unwise to intentionally write something into your enemy AI that people will perceive as being broken – either technically or in the “brain” of the enemy.
All in all, the AI seems to alternate between overly scripted or a sequence of randomly strung together actions. There isn’t that cohesive combat flow that I experience in the likes of Halo 3, Gears of War 2, or F.E.A.R. Interestingly, this makes the battles almost as mundane as the incessant walking (jogging) and driving that you have to do in the game. Where I would hope to have the proverbial minutes of terror to break up the hours of boredom, I don’t even get a rise out of the combat. That’s disappointing.
Anyway, while I don’t plan on finishing the game, I will put a few more hours into it and see if I notice anything else.

Assassin’s Creed: First Look

April 26th, 2009

In my latest episode of “finally got around to playing”, I have had a chance to mess with Assassin’s Creed (on X360). As with most people, I was immediately impressed by how visually attractive the game is. However, that has usually been a warning bell in the past. A lot of time “pretty game” has tended to also portend “stupid AI”. This doesn’t seem to be the case with AC.

While this isn’t normally my strong suit, I have to say that the animation model in the game is very well done. Even without the climbing and hoping and dodging, there is a lot of detail that made for convincing depth in the characters. Even just running forward and then rapidly changing to run the other direction prompts a foot plant and skid animation that is really kind of cool. I also like the fact that the transitions between animations are very smooth. Even just something as simple as changing from a walk to run and back to a walk (even on horseback) exposes a gradient of speed that does not have the hard-edged state change feel.

Look at me when I’m talkin’ to ya!

I also like how you can move around during the (long) dialog scenes. You don’t have to, but you can. In fact, I found myself wanting to pace during these conversations because it made it more realistic. Additionally, when your character says something while he is facing away from the listener, he will turn slightly towards him… not to directly face him, but to talk in his general direction. Think about it… don’t we all do this? How often do we actually look directly at the person we are talking to? It’s a nice touch with the animation that makes those plentiful dialog exchanges more livable.

Running, Jumping, and Climbing on Things

It goes without saying that the advanced animations are very well done. The climbing and running and jumping business is ridiculously well done. There are the occasional quirks but those are more of a problem with the control scheme (on the 360) than anything else. I especially like when I get too close to an edge and my dude has to fight to keep his balance.

Excuse me. Pardon me. Move, damnit!

There are ways that they incorporated this into the game design other than all of the building stuff, however. It was a very nice touch to have the character stagger or even fall when he runs into people. That makes the chase scenes through cities very realistic in that you do have to pay a little attention to where everyone is.

Of course, in order to utilize that mechanic, you have to pack the streets full of people going about their business. This was done very well. There are places where it is next to impossible to run through simply because of the press of people. Even walking through them can get annoying. I haven’t bothered to simply watch people to see where they are going, but you don’t get that initial impression of “random walk” that you get from a lot of other games. (One recent transgressor in this effort was when I was playing Oblivion where every person in the city just broadcasted “I am so lost!” I will write up my observations on Oblivion later.)

An even better touch is how the people react to your passing… not just physically, but verbally. If you bump into one of those jar-toting ladies and get them to drop their burden, they will give you a little lip. The same can be said for running into a dude when you are on a horse. I found myself being more careful moving through people just to avoid what should be relatively innocuous confrontations.

The agents in the game, be they friend or foe, do have a pretty obvious state-based behavior pattern. It works for this game, however. I also haven’t seen any situations where the FSM is stuck on stupid. The game mechanic has enough depth that the predictability of the guards is acceptable.

Now serving #43

I guess the only thing I can fault (although it still works well) is the “ninja fight” rules where the enemies tend to come at you one at a time. They do take turns, however, rather than waiting for the one fighting you to die before the next steps in. Again, this was matched well with the game mechanic of being able to grab people and throw them away. When you toss one dude away, his buddy will step right in. It does make the fight more manageable, however. On the other hand, there have been times when I have had more than one guy beating on me at the same time. I haven’t figured out what the difference is in those scenarios, however.

All in all, the game looks pretty good from an AI standpoint. I say that based on the fact that I haven’t noticed anything horrible about it. That’s usually a sign that things are working well. Figure that AI should be like umpires or referees in a sports game. They are necessary for the game to work smoothly, but you shouldn’t really notice that they are there.

I’m returning AC to GameFly tomorrow (although I will probably buy it because my kids are addicted to it) so it will be a bit before I follow up on this post. You can check to see if I have written anything else by clicking the Assassin’s Creed tag below.

Make sure you subscribe to the Post-Play’em feed to see more reviews and analysis.

Add to Google Reader or Homepage

Content 2002-2015 by Intrinsic Algorithm L.L.C.