Hello folks,

This thread is for discussing AAA games. Well, more specifically, it's for discussing how AAA studios have finally lost their last marble, but that much should be obvious.

I guess you can bring up the good or interesting stuff they're doing too. If you can find anything.

Middle Earth: Shadow of War (WB Games)

I really liked the first game! It had great art and music, and I thought it was fun, but somewhat repetitive by the end. Above all else, it fit at least somewhat well into the Middle Earth universe. Not perfectly, but it took the story in an interesting direction (making Celebrimbor actually help Sauron forge the One Ring, under duress, rather than just the lesser rings).

So here's the first problem: If you liked Lord of the Rings at all, this game will make your blood boil.

The main premise of the game is, Talion/Celebrimbor forge their own ring of power. Mechanically, you know, that's not a problem - in one of the Shadow of Mordor DLCs they gave you the One Ring, and it worked pretty much the way it should, if it is wielded by someone who doesn't just want to turn invisible. So it's not a problem of gameplay. It's a problem of story. The One Ring required Sauron to invest almost all of his remaining power, one of the mightiest of the Maiar. This new ring looks like the One Ring, and I've gotten the impression that it's supposed to be one, but that runs directly counter to the entire story of the Lord of the Rings. There are other dumb story things, too, like at one point you fight a Balrog. You shoot it with arrows in its glowing weak points! #wow #whoa

But then there's this, which deserves special mention. It should offend you even if you aren't an LotR fan. Sexy Shelob:



Yeah, that's right. Shadow of War features Shelob, the giant spider from Return of the King. Except in this game she's played by an attractive woman in an evening gown. Why? Because... um... ****, that's why. They could have just as easily left her as a giant spider. But they didn't!

I mean, this should piss off Lord of the Rings fans for several other reasons, too. There's a lore problem here. The developers don't seem to understand the difference between Ungoliant and Shelob, because unlike Ungoliant who possessed most of the powers of Morgoth, Shelob was just a big dumb spider with no significant magical powers. Certainly not the power to turn into a smoldering human seductress. Besides that, Lord of the Rings owes much of its popularity to its resonance with the 1960s counter-culture and second wave feminism, as one of the relatively few big works at the time featuring strong female characters who chafe against their gender and cultural roles. It's safe to say that Lord of the Rings fans are generally not going to be impressed by this kind of thing, which really makes you wonder who they think they're making these games for. Certainly not non-fans, because they're all playing Assassin's Creed.


Speaking of discrimination, they've also gone whole hog on price discrimination. There are loot boxes and in-game currencies, and you will be able to buy more of them:



And just to emphasize this point, in case you didn't watch the video above, this is a single player game. You are not even showing off fancy skins for your friends, these items you buy are only meant to help you skip content in your singleplayer game. Which begs the question: why are you supposed to pay $70 for a singleplayer game full of content that you would pay even more money to skip???

Oh, and they're going to give away loot boxes in pizza rolls.




Summary: A franchise title that Lord of the Rings fans will hate, and brings nothing that non-fans would care about. Laden with micro-transactions to an offensive degree.

The Elder Scrolls 5: Skyrim VR (Zenimax)

Described by Kotaku as a "complete dumpster fire" for, believe it or not, the poor gameplay. I know, I was surprised too. Apparently re-releasing the same 2011 title for the fourth time wasn't enough to deserve that appellation.

I guess Zenimax can't figure out how to make a good game where you stand still and look around at stuff. I am shocked.

Meanwhile, Todd Howard has recently admitted that Elder Scrolls 6 isn't in development, they have no plans to work on one, and it's not a priority for them at all. Some Forbes blog dude thinks it's because Zenimax is worried about TES 6 cannibalizing sales of Elder Scrolls Online and their endless Skyrim re-releases, which, um,... yeah, it probably would.

Thankfully I am still working on beating Morrowind. I may finish the game before I die.

If there was a gaming genre equivalent to 'dad rock', I'd be all over it.

You're already playing it?

Actually... I gave up sometime in 2005.

I don't think they thought about the implications of this.

I don't find it arousing, at all, that her vagina releases 1,000,000s of tiny baby spiders at any given time. Am I missing something here?

That Kotaku article is seriously misguided.

1) Skyrim VR is not out. They were testing a build. It's not finished. Feedback will help improve it.
2) Its main intent is not to make tons of money (I think), it is to gain experience in working with VR. Zenimax is also working on Fallout VR and Doom VR.

I am sure Bethesda will do their best because they are one of the few good guys out there, who, in my humble opinion, love video games, and love what they do. I, for one, would probably buy either Skyrim or Fallout VR. Probably Skyrim since I'd played more than 500 hours of Fallout 4 already. So yeah. I've nothing but respect for Bethesda; "you're not making the game until you're playing the game."

They are also working on 2 new games apparently, "which are bigger than anything we've done before."

As for the next Elder Scrolls, I believe Todd Howard mentioned that they will make it, they just don't know when, as the technology that they want to use in it simply doesn't exist yet.

Maybe iterating on the current cutting-edge VR will make it exist all the faster.

Who knows, right?

I trust Bethesda. Kotaku -- not so much.

Great thread, though.

Other points:

I found the original Shadow of Mordor boring. I got it a couple of years after it came out, though.

Jimquisition dude runs a tight show. Did he sew that spandex suit yet?

Let's hope the day never comes.

P.S. Did you know that a Hungarian company owns Kotaku? Gawker Media (https://en.wikipedia.org/wiki/Gawker_Media). Made by a Hungarian Brit.

Well, now you do.

Bethesda PR?

more like triple-a lames amirite

Ooh, I'll write a nice post about visual design of Diablo 3 and why so many people complain about it.

In terms of recent "commercial failures" in Triple AAA we have to look no further than Mass Effect: Andromeda. They gave it to a team that couldn't quite deliver, released early, etc. Toxic fanbase did not help. Still a not a bad game, but cracking at the seams. Or the Hitman frachise fall from from glory (although I haven't played the episodes, apparently they did not bode well for the studio).

The episodes have been widely praised as a return to form (of the Blood Money variety)... what did you hear?

Anyway, almost 14 years on and I'm still shocked that a professionally made commercial non-indie game included this level:

[Unable to find specified attachment]

I'm willing to bet that even Jedi Repair Shop has more pro quality action going on.

Jeckel_2_Dope, yo.

They've got 2 months to fix it. That's not a lot of time to fix what sounds like some serious defects.



No, an American company owns Kotaku. Gawker Media hasn't existed since Peter Thiel petered the thiel out of it.

I'm an idiot. I took this to mean you were implying that Morrowind hadn't come out yet.

But yeah. I guess I am already playing 'dad gaming'.

They weren't "testing a build," they were demoing it at QuakeCon. You don't demo something at a major game show if you're not ready to stand behind it.

It should have been obvious before the project was greenlit that "Skyrim in VR" is a pretty horrible idea. The complaints enumerated in the Kotaku article aren't the result of an unpolished build, they're very predictable consequences of the type of compromises you have to make to make a game like Skyrim playable in VR.

I'd be amazed if "Doom VR" wasn't even worse--the whole game is built around encouraging player movement, which is kind of the achilles heel for VR shooters.

Doom has the distinct advantage of being a first-person shooter, which is the only current game genre that halfway works well in VR.

Will decent AI ever come to video games? Recently I tried picking up Red Dead Redemption, and while I loved so much of the game - story, setting, soundtrack, the thing that turned me off and made me quit was how mind numbing combat was. It's simply not fun to fight fifty carbon copy AIs every single mission.

self-aware video game AI has existed for at least a couple decades

What are you talking about? FPS is one of the worst possible genres for VR. It's probably the only one I can think of than can be adapted in some way that's not incredibly awkward. There aren't any good ways to resolve the motion discrepancies between walking through a large environment and either sitting or walking in a small room. The best they can do is some kind of weird teleport or dash thing.

Any sort of cockpit game like a racing or flying simulator is perfect. Both in the real and virtual world, you are sitting down, so body motion maps directly. Make a new Rouge Squadron game for VR. That would be amazing.

You could do an RTS or platforming in room scale if you made the map a sort of virtual hologram that you could walk around and through. It's not going to work for really competitive things, but it's probably got some potential.

Yeah, Farpoint tried to address this by having maps that were essentially long corridors and preventing the player from turning (unless you turned on "advanced mode"). It helped alleviate the motion problems, but only partially, and it severely restricted the types of spaces they could build.

Doom was designed around having the player constantly moving, running and jumping around an arena with dozens of enemies. That's going to be puke central in VR.


"Decent" AI could mean a lot of things. Intelligent AI w/a complex decision-making process tends to frustrate players because it seems unpredictable or "cheap" (I remember reading about AI that got so good at distracting and flanking that players got angry that it kept "spawning" behind them).

So combat game AI is usually designed to be super simple and facilitate a given player experience. I don't remember much about the RDR enemies--maybe they could have more different behavior states?

One alternative to artificial stupidity that doesn't feel cheap is to give the player control of an entire team of walkplayers that can be cycled through, and giving the player the upper hand by providing intel about the map, and then let player plan the attack route ahead of time.

That said, the AI in Rainbow 6 / Rogue Spear kinda felt dumb nevertheless (but simply had much better aim than stormtroopers).

In a way, the bad aim and unshielded fleet of starfighters sorta makes the imperials the perfect cannon fodder for shoot `em up games.

I believe this is what you're referring to. So I'm going to speak to their comments.

For one, I don't know if the conclusions they draw are entirely valid. I think something they're missing here - and something they didn't test for - is player expectation. FPS is one of the most mature game genres. And in most FPS games, the AI is largely the same, and is largely very stupid. A big factor contributing to those player's frustrations is probably just that: they expect the AI to act differently because that's how FPS games have always been. I'd be curious to see players engage their "more intelligent" AI knowing beforehand the AI was clever, and seeing how they deal with it. It also helps to consider that the player would probably be equally as frustrated against 3 human opponents, I mean, 3v1 is bs and unfun no matter who you are.

But the problem with AI is deeper than that, IMO. I mean, take any FPS game where you can sneak up on guards. The code is basically: for some cone, probably about 90 degrees, in front of the enemy, they can see the player, and from that they have a maximum range of vision. If you're in that cone, you're spotted, else, you aren't. Games like MGS3 make that more complicated by adding camo, which basically just scales how far guards can see based on what camo the player is wearing and where they're standing. Rarely in games is it ever more complicated than this. The game simply calculates the distance from the player to the AI, and if it's close enough and a few if statements are checked, the AI goes into "aggressive" mode. Maybe advanced games will throw in some Markov chains that change between states. Or the AI will run when it's health is below 20%. Nothing really unpredictable though.

However, a human has to visually scan the 3D environment, locate and identify shapes as meaningful and act. I mean - live visual processing is in its infancy, and hardly works to keep cars from crashing with hardware beyond your average consumer, so, to expect AI will ever be at this level is silly, but, I mean, maybe there's something we can do that's slightly more advanced?

I mean, AI isn't supposed to be actually "intelligent", it's supposed to be a simulacrum of human behaviors. I would just expect that maybe someone would try to train bots with neural nets, fed input from players in a live PVP game, or something, anything that's more than this really rudimentary stuff we have. I mean, the AI in most games is basically the same as what Reverend Jones quoted. Maybe with some more if statements and a few more behaviors, but nothing revolutionary.

>>Doom was designed around having the player constantly moving, running and jumping around an arena with dozens of enemies. That's going to be puke central in VR.

This is how it looked in June: https://www.youtube.com/watch?v=EjwGztQQXmU

DOOM VFR - Virtual ****ING Reality

Starts 15 seconds in.

I guess the problem again with neural nets is, we don't have strong enough computer vision software for a neural net to play a visually intensive game like a human can. AI's basically live in vector spaces, and a neural net, if given access to basic coordinates, and told to maximize killing/HP, I'm sure could figure out how to headshot the player from max range repeatedly. Maybe AI is just destined to be kind of ****ty.

Okay, for instance: Oblivion had Radiant AI, which I think we can all agree ended up completely damn awkward:



However, that was 2005 technology, and limited. Voice synthesis software has improved quite a bit. If someone trained a Markov chain bot with good samples and really massaged things, and used a newer generation of not-completely-awful voice synthesizers, we could possibly have a Radiant AI that has much more interesting, sensible conversations. Such a thing is conceivable, and I don't think would require the world's greatest research team. But I doubt that will ever happen, I think games R&D is for the most part dead, other than VR stuff.

And frankly, the nonsense conversations of NPCs in Oblivion is still pretty interesting to me as background chatter - it's in some ways better than Skyrim, where NPCs don't converse, and instead leave you in a weird sterile world where NPCs only speak in limited contexts and say the same things constantly.

Nude walkplayer mod for Morrowind presents: interesting NPC dialog

Games can be as good as novels are, so long as the writers are funny. Bring back the adventure genre, forget this AAA crap. If I wanted to watch a multi-million dollar trainwreck I'd go watch a Hollywood blockbuster.

A controversial subject where many contradictory opinions are welcome.

First person shooter does not necessarily imply large environments, or fast paced locomotion. It implies first person and shooting. These are things that VR does very very well. VR gives you fine control over the firearms you're holding, better situational awareness, and the improved immersion of VR lends a lot to that kind of game (and generally speaking most action games). Locomotion is an unsolved problem, but if the game is designed for VR, you are given other tools to solve the same gameplay problem (for example, dodging or shields).

It's important to remember that VR isn't just an alternative display technology, it's an alternate input technology, too. You can't expect games that are designed at right angles to VR to work well with it. (This is a lesson Bethesda should learn too.)

Simulation sickness is very much a problem in sitting, cockpit based games (N.B. this is the domain where simulation sickness was originally documented). There are other problems with cockpit games, too. Most cockpit games require interaction with objects that are not tracked by VR, including your desk, chair, and HOTAS. This can be challenging or even cause injury in some cases. Standing FPS games with a clear play area do not suffer from any similar problems.

There are lots of potential applications. You actually sound more positive about it than I feel. I don't personally think RTS, platformers, or any other currently popular genre will ever transition well into VR. It is simply too different, and everybody is too married to convention right now to see how it could work.

This is where I disagree, I think--game AI is NOT supposed to be a simulacrum of human behaviors, it's supposed to be a curated set of behaviors that serves the game's mechanical design & facilitates a rewarding player experience. If enemies in most games acted exactly like real people, the games would all break.


I think the best point of comparison for this is something like online play. If I'm playing PUBG, or DayZ, or Battlefield or whatever, it's with the understanding that at any moment I could die. I might never even see the person who did it. I'd just be done. It's ignominious and often frustrating. People are willing to put up with it because the fact you're playing against other human beings makes it essentially "fair," and winning against another person feels good.

I don't think that kind of enemy behavior translates well to a single player experience. For one thing, it would hamstring the developers' ability to ratchet tension and danger up and down, if they can't meaningfully predict when and where the player is likely to die. It could interfere with narrative. It would also be more frustrating, I think. People expect single player games to be carefully crafted experiences, and therefore fair. If they're killed by an AI they can't see and don't know what they did wrong, it feels unfair. They're going to assume the developers made a mistake. On the other hand, if they DO beat the AI, it won't be nearly as rewarding, because they'll assume the AI has been tuned to be just smart enough to be beatable. There's no real winning here.

I definitely think there's space for AI behavior to be more interesting. But what makes it good is going to be wildly different from project to project, and it always has to be w/an eye toward what the player sees.

I agree, also, that there's not enough R&D for this type of thing. Commercial studios don't want to devote the time and money to just fooling around, and academic people (like the UCSC games program) have the interest in AI, but not the talent or work ethic to produce anything valuable. It's a bummer. I'm still hopeful we'll see interesting stuff from somewhere, though, w/Unity and Unreal democratizing game development.

To echo Thawn's point:
https://www.youtube.com/watch?v=1FBGR6vmNeU

Right - games like CoD, where one person fights entire armies, would be impossible. my idea wouldn't be a shoe-in for all game types. That doesn't mean it wouldn't have a place in any game.



I wouldn't argue the success of PUBG has anything to do with the fairness, the game is patently unfair. How many times do you drop in near an enemy - their house has an AR, your house has nothing, and you get gunned down? I believe the main reason PUBG is so successful is, you see a very wide variety of player behaviors, you encounter a wide variety of combat. And, it's also a game that rewards creative, out-of-the-box thinking. My most memorable PUBG moments are all quite different from each other.



So I agree it can be not enjoyable for a player who cannot beat the AI - I don't see the converse being true though, I don't think every player will just assume the AI was tweaked to be easier. Again, I also think this wouldn't work for every type of game.



So - yes, complex doesn't in itself mean good. I think complex AI needs to be carefully managed to make sure it's not destroying gameplay experience. In response, I'll tie it into a response to what Gebohq posted, his number two point: "Do the behaviors create interesting decisions for the player?" I would argue that, in the vast majority of AAA games out there, the answer is an emphatic no, that AIs are routinely predictable and boring, and fighting them well comes down more to tedium than interesting decisions. So - yes, more interesting, and I think that, in many cases, we have exhausted what is interesting with the limited scope we have for AI, and we need to expand that scope to be able to expand the behaviors AI are capable of.

Hopefully, like you said, democratization of game development can lead us somewhere better.

It's easier to make an AI that can see everything in the map and has instantaneous reaction time and perfect aim than it is to make one that's more limited, so all AI is "tweaked to be easier" by nature. I didn't mean to suggest that inherently makes it not fun--but there's a psychological component to beating a human player that (we assume) is trying their absolute hardest to beat us. At least for me. Pwning a human is more meaningful than pwning an NPC. Although I suck at games and haven't played Dark Souls, etc. Maybe there's exceptions there.


Agreed 100%!

Do these non-indie "big name" VR games actually make money or do they get kickbacks from VR gear companies? The amount of people with active VR gear (and willingness to pay for these games) really can't be that high, or am I vaguely underestimating that number now?

Granted that map was terrible but I feel like individual professional mappers/level designers had much more agency back then than they do now. So there were a lot of memorable levels/maps along with more fun oddballs when there were just a few guys focused on each map (like a singular vision). A type of deathmatch level that would resemble one of JK's more memorable almost-confused-in-approach MP maps would, for better or worse, not "show up" in today's landscape.