265 replies to this topic

[Toothless]

Its weird because despite a lifetime of mechwarrior games and love of battletech, starting in 1995 with MW2, I see all of that and still dont want to give PGI any money. It's almost as if over seven years of mismanagement with MWO cant be resolved by screenshots of looks-good-for-2015 graphics, hyperbole, and a bad HUD.

[Nightbird]

Well, let's see if the reviewers agree

[Novakaine]

Wonder if there will be a campaign or scenario mods.
I'd take every one my Battletech scenario books and mod the scenario's.

[YueFei]

Jyi, on 03 December 2019 - 02:16 AM, said:

No, personally I just want AI to play like a sane entity, not like a drooling cavetroll with a split personality. It DOES NOT NEED TO BE geared around a casual player with a joystick. This game is supposed to sell to other people besides mech dads and other hardcore fans. How many times does it need to be repeated: it is BAD GAME DESIGN to make the game for the lowest common denominator.

If you want people with joysticks to be able to play, you make a good AI and then add difficulty options that make it easier to tackle, not the other way around.

I've explained this multiple times, but instead of any argument to prove me wrong, you all just keep ignoring me.

I have to agree with this. You don't even necessarily need explicit difficulty settings. Wing Commander 2 is an example of a game where the AI dynamically adjusted itself based on the player's actions. For fun, you could of course still hard-set the difficulty, just to see what would happen if the game didn't "hold your hands", so to speak, and when you did that, a single AI fighter was dangerous enough to kill you in a single burst of fire whilst pulling a skidded attack, and you'd have to work your butt off to evade and counter.

If a game made in 1991 can sense a player's competency and dynamically adjust the AI, why can't a game made in 2019?

[Jackal Noble]

Toothless, on 03 December 2019 - 06:12 AM, said:

Its weird because despite a lifetime of mechwarrior games and love of battletech, starting in 1995 with MW2, I see all of that and still dont want to give PGI any money. It's almost as if over seven years of mismanagement with MWO cant be resolved by screenshots of looks-good-for-2015 graphics, hyperbole, and a bad HUD.

No doubt. In my history of posts (if you can stand reading witless diatribe) you'll see that I have never been easy with this company or it's previous product because of what you mentioned. There are quite a few current/former MWO players that suffered burnout because the game never changed, had many minimalist aspects that seemed to be foregone rather than fleshed out. This is known.

I am the last person to recommend the game, I have been very critical of PGI. Despite that and being in the wave of anger/betrayal at the EGS deal for this game, and prior to that years of MWO bungles. Regarding the switch to EGS, the more I think about it, the more it makes sense - Using unreal engine, why wouldn't you use the Company that makes your engine and -not sure if this is true- gain access to assets. Was it handled well, not really, but in hindsight at that point we were frothing at the mouth for anything to spout anger about.

We have it so good, today, where you can use multiple platform to download game(s) to the point that there is some serious entitlement going on. I could go on forever about that, but I'll leave at, it's a pretty ****** up world, where players feel that they are on the same level to 'demand' development or become offended at a digital good or practices. Don't like it? Surely, you can find another game in the galaxy games that are already at a point where you can not realistically play all of the best games that come out every year in one lifetime.
You don't have to be convinced or otherwise, this was just to put out some info and my personal opinion about what I've experienced thus far from MW5.
I am very surprised, both pleasantly and expectantly from what I've played.
Bruh, the graphics are really nice and well-optimized overall (belch flamer). They are contemporary and the mechs and vehicles etc look awesome. The graphics look 2019-20.
I included screenshots because I like 'freezing' the action and using the camera to pan around the scene. It's cool as hell.

Think of my op as a post by someone who has felt very similar to you in terms of product development. I'm thinking the big difference might turn out that it turns out that PGI is better at making a good single player game, after having years of forum supplied fan content - allowing them to see what elements would indeed make for a successful venture. Also, having the benefit of witnessing how parts of HBs Btech was received.

[Jackal Noble]

YueFei, on 03 December 2019 - 10:28 AM, said:

I have to agree with this. You don't even necessarily need explicit difficulty settings. Wing Commander 2 is an example of a game where the AI dynamically adjusted itself based on the player's actions. For fun, you could of course still hard-set the difficulty, just to see what would happen if the game didn't "hold your hands", so to speak, and when you did that, a single AI fighter was dangerous enough to kill you in a single burst of fire whilst pulling a skidded attack, and you'd have to work your butt off to evade and counter. If a game made in 1991 can sense a player's competency and dynamically adjust the AI, why can't a game made in 2019?

You realize that the demo, is just 5 instant action maps with pre-defined difficulty? In the demo, you don't even run into any full-assaults just some 80-85 tonners and they are sparse intermixed. The game will have all sorts of settings (similar to your weird WC2 analogy).
The AI does react differently based on your vector of attack to your objective.
Difficulty - define difficulty. Is it running a Commando into a drop that recommends 300 tons? I can guarantee you won't finish that mission.
Is it having a modifier- where you can enhance the enemy accuracy/armor ratio?
Is it feeling like you are on a 1-4 mech lance fighting off an outpost or city armory?
If you are alluding that AI in a game that is 30 years old is on the same level of complexity then I am going to need to sit this one out, because there is no amount of logic that I could communicate for us to agree with each other.

Edited by Jackal Noble, 03 December 2019 - 11:13 AM.

[Jackal Noble]

Novakaine, on 03 December 2019 - 08:28 AM, said:

Wonder if there will be a campaign or scenario mods.
I'd take every one my Battletech scenario books and mod the scenario's.

I know there are a few groups that were working on some lore-based scenario mods. Haven't kept up with it, but would assume that there will be some of that for sure.

[Koniving]

Wait, so are they spending time and effort patching the demo, or did you mean they patched the game we'll soon get?

[Koniving]

Novakaine, on 03 December 2019 - 08:28 AM, said:

I'd take every one my Battletech scenario books and mod the scenario's.

I'd enjoy playing them. Many basic scenarios could easily be played. I'm a bit more worried about the really awesome ones as they'd be a lot less simple and may need handcrafted maps.

(On a side note, there's generated and handcrafted maps. The ones we get in the demo are likely handcrafted when you consider it's the same map each time for each of the 5 maps. As of right now we don't know about the quality of generated maps, but I imagine more complicated scenarios that may need some scripting would be difficult to pull off in procedurally generated maps.).

[Jackal Noble]

Koniving, on 03 December 2019 - 11:24 AM, said:

Wait, so are they spending time and effort patching the demo, or did you mean they patched the game we'll soon get?

The game for certain.
However, I'm pretty sure the demo difficulty was enhanced. I had a 1.5 gb update and after that have noticed a sig spike in difficulty, where before I would just mow down and could solo pretty well. Now, I'm throwing together teams at the recommended tonnage and enjoying how it plays out, usually finish but sometimes get taken out. It's fun.

Edited by Jackal Noble, 03 December 2019 - 11:30 AM.

[Koniving]

This is why it's puzzling that they'd update the demo, but I suppose if they didn't have to change much (like update a class file).

[VonBruinwald]

BlaizerP, on 02 December 2019 - 07:34 PM, said:

In what world did this happen?

MechCon 2018. It was behind closed doors though. It's been confirmed that the, Gargoyle and AC2 received "tweaks" as a result of this meeting, possibly the Piranha but that was attributed to non-specific feedback. There was also some other tweaks to 'mechs weapons but hasn't be confirmed as a result of the meeting, most likely it's just PGI's dartboard doing its thing.

[YueFei]

Jackal Noble, on 03 December 2019 - 11:10 AM, said:

You realize that the demo, is just 5 instant action maps with pre-defined difficulty? In the demo, you don't even run into any full-assaults just some 80-85 tonners and they are sparse intermixed. The game will have all sorts of settings (similar to your weird WC2 analogy).
The AI does react differently based on your vector of attack to your objective.
Difficulty - define difficulty. Is it running a Commando into a drop that recommends 300 tons? I can guarantee you won't finish that mission.
Is it having a modifier- where you can enhance the enemy accuracy/armor ratio?
Is it feeling like you are on a 1-4 mech lance fighting off an outpost or city armory?
If you are alluding that AI in a game that is 30 years old is on the same level of complexity then I am going to need to sit this one out, because there is no amount of logic that I could communicate for us to agree with each other.

You're talking about completely different things, which are completely besides the point. Dropping lighter or facing heavier foes has nothing to do with AI behaving more or less competently based on player competency. That's a matter of player choice in mech+loadout and map/mission design. Wing Commander 2 didn't increase difficulty by spawning more enemies, or heavier fighters, it increased it by making the AI fly smarter and harder.

The fact that you think your red herrings are relevant to that line of discussion is entertaining, but useless.

[Koniving]

VonBruinwald, on 03 December 2019 - 11:56 AM, said:

MechCon 2018. It was behind closed doors though. It's been confirmed that the, Gargoyle and AC2 received "tweaks" as a result of this meeting, possibly the Piranha but that was attributed to non-specific feedback. There was also some other tweaks to 'mechs weapons but hasn't be confirmed as a result of the meeting, most likely it's just PGI's dartboard doing its thing.

If you know the real reference in PGI, or seen one on someone's dashboard in MWO, you'll know MechCon isn't the only place.

[Jackal Noble]

YueFei, on 03 December 2019 - 12:13 PM, said:

You're talking about completely different things, which are completely besides the point. Dropping lighter or facing heavier foes has nothing to do with AI behaving more or less competently based on player competency. That's a matter of player choice in mech+loadout and map/mission design. Wing Commander 2 didn't increase difficulty by spawning more enemies, or heavier fighters, it increased it by making the AI fly smarter and harder.

The fact that you think your red herrings are relevant to that line of discussion is entertaining, but useless.

You're right, I'm talking about Mechwarrior and your off talking about Wing Commander 2.

[Koniving]

YueFei, on 03 December 2019 - 12:13 PM, said:

Wing Commander 2 didn't increase difficulty by spawning more enemies, or heavier fighters, it increased it by making the AI fly smarter and harder.

Indeed.


Here's a good AI design.
For skirmish battles, they built the AI to compete with and be able to defeat competition players at legendary difficulty, and the AI is stripped gradually with each lowered difficulty.
For the campaign AI skills are unlocked over time, ramping up with your own skills (showing exceptional skill unlocks them sooner and it's possible to unlock all the skills of the AI long before they're scripted to be at the latest).

If the AI is built for lower or middle people, it doesn't really go up which is the point that I think Jackal is missing.
And the difficulty in MW5 isn't intended to be ramped by your settings so far as I can tell as there is no option to change difficulty anywhere in the options.
As such the difficulty is increased by the number of enemies you face.
Because they definitely are not getting smarter. They have no comprehension of teamwork and the AI literally see each other as moving obstacles.

And what is the AI programmed to do with moving obstacles, PGI? Lets let PGI answer that.
"The AI would destroy their own base trying to kill the player." "We didn't want them to do that but we kinda liked it." "So we made it so that if there's a building [an obstacle] between the unit and the player, sometimes, the AI will blast its way through."

So the AI is literally programmed to destroy obstacles -- sometimes -- in order to get to the player.
The units see each other as obstacles.
They're PROGRAMMED TO KILL EACH OTHER...sometimes.
THAT is the biggest issue I have with their stupidity.

Overall my personal issue is that the higher the difficulty setting not only produces more enemies, but because there are more enemies they hit each other more and show their acute lack of awareness of each other (two lights running into each other isn't as funny when there's no collision and both try to turn in the same direction to get around each other, meanwhile one is shooting THROUGH the other to try and hit me).

But further to that, while true they've certainly become more aggressive than the 1.1b that I originally had, they still fire less often when there are more of them. Whether intentional to give you a chance or by fluke is unknown, but in doing so it opens a number of exploits that one can do.

The AI overall also has no concept of threat assessment, this is evident by the fact that AI does not react to its friendlies trying to destroy it. This goes for the enemies and your ally wingmen. They do not recognize threats that do not appear as "red" to them.

Mechwarrior 2 mercenaries had wingmen that could recognize You as a threat if you attacked them.
If you trick Mechwarrior 2 mechs into shooting each other they will engage each other (this is why you usually only saw one or two at a time).

In Turok, a game on the N64, units of different species/classes will engage each other if they are attacked on accident by another one. Case in point if both a Poacher and a Raptor are attacking you and the Poacher hits the Raptor, it's ON! And a Priest versus a Purlinn is a fight to see. Level 4 if you activate all the enemies at once you can get a brawl going in which there can only be one victor out of 8 enemies.

Mw5 has no concept of this.
As such, it likely has any "friendly" units not supplied by you on scripts because it's highly doubtful the game could handle the concept of 3 teams.

Jackal Noble, on 03 December 2019 - 12:28 PM, said:

You're right, I'm talking about Mechwarrior and your off talking about Wing Commander 2.

He's making a comparison by pointing out a game that has an AI that ramps up difficulty by playing better, not one that thinks more enemies that get dumber = more difficult.

Edited by Koniving, 03 December 2019 - 12:54 PM.

[Jackal Noble]

It was a casual reply (non-serious). When someone, in the archive of all game A.I. in history decides to pull out WC2 A.I. as their counterpoint it's obvious that there's not actually going to be a discussion with said person.
Enjoy something for nostalgia purposes, but don't come here with that and think anyone should take you seriously -

The A.I. displayed in WC2 is definitively inferior

Also, comparing a six axis space shooting game to a land-based mech game is apples to oranges from many standpoints.

[Tordin]

FupDup, on 02 December 2019 - 07:30 PM, said:

Bruh, please. People are gonna pretty quickly figure out the objectively strongest mechs/guns/strats just like they did in TT, HBS-BT, and every MW game. I dunno how you think that never having patches will prevent some guns or robots from being the "meta."

Only solution I see is making every option viable, just in different situations. Alternate "upgrades", role for specific situations if you like.

Like with the case of the AC/ LBX 10 family in MW5M now. AC 10 standard seems to be universylly less stellar than AC 10 Burst, LBX 10 ( scatter/ shotgun ) and LBX 10 Slug.

The burst AC 10 can stand up to the LBX since it have a fast recycling time as it seems, no I havent tested it yet. A good infighter version. Its not scattering like the LBX 10, its not throwing a singel slug like the AC 10/ LBX 10 slug but hold its own it seems, like a compromise.

Also once modders for EGS and soon, Steam gets to work and sinking their fangs into the files to the fullest. Blance and options is of least concern.

Edited by Tordin, 03 December 2019 - 12:54 PM.

[Koniving]

Tordin, on 03 December 2019 - 12:54 PM, said:

Only solution I see is making every option viable, just in different situations. Alternate "upgrades", role for specific situations if you like.

The issue with a game that has an upgrade system to weapons, is like in HBS BT once you get a level 5 AC/5, that level 1 AC/20 is worthless.
That's why my mod aims to make every weapon different but equal, catering instead to variation and playstyle.
For example the ancient Diverse Optics type 2 Medium Laser (of Wasp fame) is canonically with an issue of lower range but preferred for its wallop within that lower range. So it is between 5.5 and 6 damage, split into two shots of 2.25 or 2.5 damage, at the sacrifice that the damage starts reducing gradually after only 180 meters. This gives a better punch for those close in skirmishers and the weapon counts as a Small Slot. (Last edit I swear: This isn't a trait that goes across all Diverse Optics, just this specific laser as this is one of the two oldest ones still in production. The Diverse Optics Type 18 ML is one of the three that set the "standard" for medium lasers with the other two being slight variations on the same premise.)

Meanwhile, Intek, known for exceptionally longer ranges within a given laser class and the longest ranged Large Laser before the release of the Diverse Optics Sunbeam that generated the new "ER" classification (and comedically, the Intek Large Laser and the DOS ER Large Laser have the same range)... What dethroned the Intek Large Laser and makes it not count as an ER Large Laser despite matching the range, is Intek LLs require twice the number of shots as the standard large laser to deliver its damage. So if the standard needed 3 shots, the Intek needs 6. If the 'standard' needs 6, then the Intek needs 12. (Just examples.) There's no time difference because Intek has a fast enough firing rate to get the same damage in the same time, but it requires more skill to land those shots at that exceptional range. (Every weapon must be able to reach rated damage within 5 seconds or less, and because this would make time to kill significantly longer health is going to 1x as even then mechs are likely to tank a lot more firepower than you might expect given the non-front-loaded nature of all the weapons) (Overall this goes across all Intek laser classes.).

So while that specific Diverse Optics laser caters to close in light brawling, the Intek laser is ideal for someone that prefers to stay further away from the action to sharp shoot from ridges and the like.

Where I think the mod would make the game more interesting, is these weapon variants are also going to be supplied to mechs as canonical, including sub-variants (for example the Kali-Yama Hunchback 4G and the Komiyaba/Nissan Hunchback 4G are extremely different in their weapons loadout with the Kali-Yama sporting more aggressive lasers but a significantly weaker burst-fire AC/20 which does 1.67 damage per shot (bullet) in four 3-round bursts to get 20 damage (TRO 3039, these start rolling out in 3030), while the Komiyaba/Nissan Hunchback 4G is sporting an AC/20 that can rapidly spew a 5 round salvo of the second-highest caliber in the Inner Sphere with shells that can blast holes THROUGH mechs.. but has weaker, faster shooting lasers (TRO 2750, 3025, these are the ones in the field though a non-modified 4G sporting the 180mm Tomodzuru is actually rare, as most become the sub variants like 4P, 4J, 4SP, etc).

Edited by Koniving, 03 December 2019 - 01:30 PM.

[MischiefSC]

Jyi, on 03 December 2019 - 02:16 AM, said:

No, personally I just want AI to play like a sane entity, not like a drooling cavetroll with a split personality. It DOES NOT NEED TO BE geared around a casual player with a joystick. This game is supposed to sell to other people besides mech dads and other hardcore fans. How many times does it need to be repeated: it is BAD GAME DESIGN to make the game for the lowest common denominator.

If you want people with joysticks to be able to play, you make a good AI and then add difficulty options that make it easier to tackle, not the other way around.

I've explained this multiple times, but instead of any argument to prove me wrong, you all just keep ignoring me.


It's pretty bad, yes, but definitely not worse than MW5. Not even by a long shot.


Seriously? You're saying a 100% cosmetic thing that can EASILY be modded in (one of these "low hanging fruits" of Russ) is more legit than a gamebreaking base mechanic like AI?

Seriously?

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.

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).

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.

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.