MischiefSC, on 03 December 2019 - 03:01 PM, said:
Your assumption about game design for the majority of consumers for the product doesn't become valid because you call them "lowest common denominator". Especially for an otherwise niche product accessibility makes or breaks you.
Mech dads are not the "majority of consumers". Unless you decide to make a game SPECIFICALLY for them... and then for some reason don't even bother implementing proper joystick support. LMFAO.
There's plenty of people like me who want a good Mechwarrior -game with proper challenge, and more people who didn't know they wanted a good Mechwarrior -game with proper challenge, but could realize they do when such a game is released. Sadly, that probably won't happen now, because PGI truly seems to be clueless.
Quote
Your assumptions about game AI are totally divorced from reality. Of COURSE it's possible to design an AI to do whatever. The issue with game AI is *always* 3 things. Investment of time/available resources (coders), game resource cost (every single decision or call an AI makes reduces your ability to improve graphics and/or total number of AI in play, allied and enemy, at any given moment) and flexibility (does the AI require hand placement and map design resources to work, adding effort and cost to level design and hurting modability or are they just spawn and go, also for MWO with each mech having variable range, speed, JJs, weapon ranges, heat profiles, how flexible is it for the individual mob).
Divorced from reality? Because I want an AI that's not braindead? Seriously?
Also, graphics and AI are generally done by different people or departments. You don't hire a code monkey to do graphics - or vice versa. Thus, graphics are generally not affected by AI design. However, I'm not certain how much procedurally generated maps require coding. If they'd skipped doing their stupid and pointless procedural maps and gone for a more rudimentary system, they might've had time to develop a better AI. Or they could've hired some extra people to work on the AI. Seeing as AI is one of the core aspects of ANY game.
Also, they could've concentrated on AI design for release and for example worked mod support in in a post-release patch - not the other way around.
Quote
MW5 has destructible assets and widely variable terrain and it's a shooter. Every other game out there has AI about like MW5. Most make it look better by tricks with level design (DOOM is a corridor shooter that doesn't feel like a corridor shooter) or by tweaking the mobs in ways MWO can't - damage/health buffs, weapons/moves the player doesn't have that let them exploit the unique environment they spawn in.
Oh yea, I forgot the stupid and completely irrelevant destructibility. Sure, it looks cool for a whole lot of 5 minutes until you realize how ridicilous it is that you can walk through buildings like they were paper - except some buildings that only have an outer layer of destructible parts and become completely and utterly invincible once you get through that.
Also, no, most games don't have similar level of AI as MW5. They have design that is decades more advanced. And while I detest your ridicilous argument that Doom AI is good only because of level design (I'm assuming you're talking about the new Doom -games, not the original one from -93 - though even that probably has better AI than MW5), PGI could've done that if they'd taken a more handcrafted approach to their design instead of the extravagant procedural generation - which they clearly were incapable of pulling off competently.
Previous Mechwarrior -games had more handcrafted design, and thus they could have appropriately tuned design. But nooooo, not PGI.
Quote
UI is worth criticizing specifically because it IS low hanging fruit. Sure there's room to improve the AI a bit but a stable of the greatest AI designers in the world still won't make it significantly better without a sea change in gaming hardware and/or software platform. Yeas from now the best improvement you can expect are small tweaks that'll help a bit and be very labor intensive to do. The UI however can/should see huge improvements for relatively little effort.
I think anything in the game is worth criticizing, but there are things that should be criticized more and things that should be criticized less. If you think a competent developer couldn't make a competent AI in UE4, you're clueless. I don't want to go to ad hominem -insults, but I don't know how else to say it. You're completely clueless if you think an average developer that actually put some effort into it couldn't make a LEAPS better AI in UE4.
And if the UI is so little effort, you go mod it out. I've criticized the UI since the alpha and beta -videos, but to me it's still comparatively inconsequential if the BASE GAME IS BROKEN.