[1.6.0] The game feels entirely overtuned/general feedback

[Micbran]

idk what you guys are talking about — every change was a positive change and the game is vastly improved, you just need to adapt.

— signed, someone who doesn’t play 1.6

[Snarvid]

Thank you Mankeli, I appreciate the granularity of your comments.

I don't have your gameplay hours or versatile win table, but I'm both Snarvidand Snarvidesque(yeah, it was a rough transition to Steam) and Steam says I have 1200+ hours of playtime. So... some basis of experience.

I would also say, re: intelligent enemies - if you're going to have your enemies be higher level and vastly more numerous than your players, they also shouldn't be smarter than them or even close to as smart. e.g. The Deepmind Starcraft 2 AI has limits on what it can do in terms of mouse APM and it's a grandmaster currently, if you took the speed limiter off it would dominate everyone, and if you buffed its units significantly there would be no contest.

But, hey - I wasn't there for the Discord discussions, so I lack the context to see how heated things got on both sides. I hope we, the experienced players, will get some but not all say in the future of this game, or for me it will go the way of Stellaris - patched by a strong developer vision that led it into something I have no interest in.

[Mankeli]

Thank you too Snarvid. I've talked to some ppl in discord and after seeing your post I decided to finally finalize mine.

Things weren't that heated in discord and that quote about making very npc smart was probably said jokingly at least to an extent. It just really frustrates me how people have made posts about the potential pitfalls of "smart AI" literally months before the release and now to see many/most aspect of it come into fruition.

Edit. and yeah I've watched some videos on humans vs the deemind ai in starcraft, it's fascinating stuff.

[richardhawk]

The omniscient AI is definitely a thing. It can phase door on you through multiple walls after you enter a level stealthed, or unerringly continue to shoot at you after you drop a high level Cloak Gesture AND move away. It invalidates a lot of tactical play considerations.

As for the other concerns, they are all fundamentally tied to numbers which can be tweaked, but I believe more issues will always continue to pop up until enemy talent level scaling is reduced. Talents were simply not built to make sense at level 9 or 13, and something is always going to be devastatingly busted as such.

[Snarvid]

Talent level scaling varies wildly, though - in most cases the majority of the return on any given talent is frontloaded into the first couple levels and then you get diminishing returns with each additional level. Weapon Mastery, for example, gives a 34% damage bonus at talent level 1.3, 76% at 6.5, 119% at 16, such that more than 25% of the top-end value is found in the first 8% of the talent levels. So I'm not sure I diagnose the level scaling as the problem in most cases (War Hound and maybe a few others excepted), it's what the talents actually do. It's unclear to me whether spell damage has been boosted per talent or spellcaster enemies have received large bonuses to spellpower or magic stat, but it's significantly more than it used to be in 1.5.

Given that darkgod said they did not recollect adding a boost to talent-based damage on the top of the second page of this topic, it's possible that this was not intended. It's equally possible that I simply said it wrong, and that only spells and not the whole subset of active talent damage had been increased and so my comment did not trigger the relevant memory (also, I rather suspect that there is a lot on darkgod's plate right now, given some of the literally game-breaking bugs discovered in 1.6, so they might not have spent as much time on their response as they would have another time, making the memory theory a little more likely). As a player, it feels like a very different set of concerns if the current model is working as intended than if it's an unintentional byproduct of other changes, with the former being much less likely to change than the latter.

Agree on omniscient Ai though. Was very sad when in testing my 200+ Stealth power cloaking device character was routinely chased down and accurately shot at by enemies 20 levels lower than him. Like a bad superhero vs. superhero argument between kids - "I shoot you" "No you can't I'm invisible" "No I have invisibility-seeing glasses!"

[RandomKesaranPasaran]

Might as well throw in my two cents and idle speculation.

I haven't looked over every single enemy but I know bosses at least are getting 1.3 categories now, which I'm pretty sure wasn't the case before and can't be helping things, since even from the perspective of a more lower-end player like myself the game has enough rocket tag issues as is (to the point where I honestly was writing off the notion that spell damage was buffed for a while, like yeah Vor was killing me a couple times but Vor also got Elemental Surge so I shrugged and moved on), so stuff like near-4k Elandar Freeze crits happening on Insane isn't surprising considering that talent's problematic scaling, even if it's the sort of thing we all had hoped ToME had moved past. I could be wrong on that point though, I kind of fizzled out on the game for a long time so I'm still a bit hazy as to what was a 1.5 change and what was a 1.6 change.

But the broader point is that it's not how much value can be found in individual talent levels or even necessarily talents with problematic scaling like Freeze. It's that, even with diminishing returns in play, you hit a point where your only realistic option is to dehumanize yourself and face to slot machine, and if rares/uniques/bosses are having their talents boosted by getting 1.3 categories, well that in conjuction with the apparent pushback against the power creep that's seeped in over the years along with the omniscient AI would explain a bit since it doesn't look like their raw talent levels were lowered to compensate, and that definitely seems to be what's happening with Insane at the very least, and it does seem to be trickling down to the lower difficulties as well, if to a more manageable degree. Still worrying, though.

[Mankeli]

Bringing up to power creep is a good point. I'd actually like the game being harder on insane than it was in 1.5 but if this was some partial design goal it really falls flat in execution. Let's take that teleporting/phase dooring monster that escapes to heal for example. If you can tank a boss 15 times before it teleports away to heal, is that 16th time going to make a difference in your survival? And the player will approach fights with the primary intention of not dying instead of killing. "Burning through everything and throwing caution to the wind" isn't really a feasible primary strategic in a roguelike if you are looking to win since eventually, and prolly soon too, you will be punished for it and your character dies.

Ofc it's not always possible to save that teleporting monster for last or kite it in peace but on many occasions it is. On those situations it's simply an exercise in tedium instead of exercise in skill or tactics. You will first guarantee that your character lives and then at some point you manage to get enough disables off or the enemy does his talent usage in "wrong" order and eventually the monster will die or you will abandon the fight. And this is not some theoretical analysis either but based on actual 45-90 min fights against bosses that I knew that I could kill eventually without any risk when they fail to escape/heal/whatever properly. I don't understand why devs of any game would celebrate wasting players as some kind triumphant moment for the AI because it is quite the contrary: it's the player outsmarting the AI and just waiting it for to fail.

On some some other occasions the new AI actually increases the difficulty instead of just increasing tedium but as already discussed, omniscient skynet that somehow aggroes on other rooms and somehow manages to teleport right next to you isn't very interesting either. Dfell stair ganks where you have to fight 4 randbosses before you even got to move away from the stairs was already a thing in 1.5 because of some banshee aggroing 1/3 of the level or something similar while it make it way to you from the other side of the level. If you add a madness difficulty style hunted debuff then yes it will make the game more difficult. But it can also make some levels unwinnable. The current situation isn't as bad ofc but does feel unfair at times. You could also add a 1d100 dice that that is rolled every ten turns or so that kills the player instantly if it hits 100. That too would make the game more difficult but it would not make the game more interesting. Uh, reading this again I kind regret posting this dice idea already, please don't implement this in 1.7 it was just a joke and is horrible.

And from the dice example we get to the random once shots. It's too early for me to say with utmost certainty if getting one shot is now more or less frequent but it feels that they are indeed more frequent now. Ofc there are probably biases in play and if you wore light armour in 1.5 without having talents that actually reward you from doing so, or had AoS or played something like an archmage then you'd gotten punished for it very often too. But melee still isn't completely a non-threat so you still want to wear heavy/massive when you can in 1.6. But like, if it the final fight is down to either elandar targetting aeryn/glitching itself with escape ai by spamming useless stonewall constantly/being maybe silenced/invisible the whole fight or alternatively potentially doing 7 K dam in one round it should be obvious that something is very very wrong.

Edit. Also forgot to mention that both the nominal increase in difficulty that is de facto just added tedium and the actual difficulty increase serve to make the game effectively longer. In prides even the monster density is increased to add to this further. Some other changes lessen this but the game does seem to take even longer to win now.

[RandomKesaranPasaran]

Bringing up to power creep is a good point. I'd actually like the game being harder on insane than it was in 1.5 but if this was some partial design goal it really falls flat in execution.

Yeah, it's important to remember why some of the power creep existed in the first place and in some places it definitely feels like some of the pushback was either misaimed or soured badly when mixed in with other changes. And yeah at the end of the day the game is still a roguelike so you expect a certain level of things not going the way you want them to but there's still a difference between a healthy level of challenge and, well, stuff like this:

Ofc it's not always possible to save that teleporting monster for last or kite it in peace but on many occasions it is. On those situations it's simply an exercise in tedium instead of exercise in skill or tactics. You will first guarantee that your character lives and then at some point you manage to get enough disables off or the enemy does his talent usage in "wrong" order and eventually the monster will die or you will abandon the fight. And this is not some theoretical analysis either but based on actual 45-90 min fights against bosses that I knew that I could kill eventually without any risk when they fail to escape/heal/whatever properly.

I would consider myself a mediocre-at-best player because I'm lazy, sloppy, and far too prone to snap plays and flying by the seat of my pants but at the same time I would like to think that I've played this game for long enough that I'm no stranger to how it actually works and as such readily believe this, because I've seen enough shades of this jumping between Normal and Nightmare in the game's current state (as well as watched/listened to parts of another person's own Insane run) that envisioning how such a fight would go on Insane ends up filling a lot of the "okay where did the Greater Mummy Lord teleport off to -this time-" drudgery, and as such wind up finding myself in this weird spot where liking the game and wanting to put in the actual effort required to be good at it are at odds with one another.

So yeah, mostly I just wanna thank you for your posts here since it's helped me get my own thoughts on the matter into some semblance of order.

[richardhawk]

Talent level scaling varies wildly, though - in most cases the majority of the return on any given talent is frontloaded into the first couple levels and then you get diminishing returns with each additional level. Weapon Mastery, for example, gives a 34% damage bonus at talent level 1.3, 76% at 6.5, 119% at 16, such that more than 25% of the top-end value is found in the first 8% of the talent levels. So I'm not sure I diagnose the level scaling as the problem in most cases (War Hound and maybe a few others excepted), it's what the talents actually do. It's unclear to me whether spell damage has been boosted per talent or spellcaster enemies have received large bonuses to spellpower or magic stat, but it's significantly more than it used to be in 1.5.

I'd like to push back against this some. It is not only effectiveness numbers that go up with humongous talent levels (though 30% is nothing to sneeze at, especially with things like armor, damage reduction, resistance penetration). Durations, radii, % chances and so forth also often go into stratosphere, turning talents into something far from their original design. I believe there are other methods of making enemies more powerful to compensate if scaling back tLv scaling would take out too much of their bite.

[Snarvid]

Again, depends on the talent, but typically duration does the same thing - the initial points get multiple round of duration, and then later ones drop off to 0-1 level improvements per effective talent level increase. There are still cases where this matters (overly strong defensive talents where uptime approaches cooldown) but I would say that would not be the problem if the talent were not overly strong initially.

My other point of skepticism on talent level mastery being the issue is that I certainly fought enemies in 1.5 with talent levels in the 11-14 range and did not have this kind of trouble.

Unrelated point on the AI change: in a narrow hallway enemies now prefer to walk away from tough, immobile blockers like the Summoner’s Jelly and try to find alternative routes to the player. This robs the Summoner of their best tool for recovering EQ unless they’re going to be hitting it themselves, and if they do Through the Crowd (not my favorite, but a prodigy designed for Summoners) their ability to do this drops to hitting it with up-close attacks (and it is questionable whether Through the Crowd allowing even this is an oversight based on the reading of the Prodigy).

[richardhawk]

The talents aren't overly strong when they are at their intended level though (SURGE OF POWER aside, which is obviously not meant to have HP pools of randbosses to scale off of). Scaling arbitrarily upward can never be failsafe nor futureproof.

I do have to say I was genuinely impressed when I saw a pack of luminous horrors pincer attack me to cut off kiting routes. I like that part of the AI.

[Snarvid]

Again, that does not account for it not having been the same problem in 1.5. An extra 0.3 talent mastery isn’t that big a deal when endgame enemies were rocking 11-14 already (believe you me, as an avid Possessor player I watch enemy talent levels more than most players). And to my very limited coding and trash-tier math skills, the fact that effect scaling for almost all skills include square roots (as in, Talent Factor = (0.2 + 0.8 * √ Talent Level) / (0.2 + 0.8 * √ 5)) does limit the impact of higher talent levels, such that it’s e.g. a 33% increase in damage to go from eft 9 to eft 16. The difference between 1.0 and 1.3 mastery is smaller than that, and that is not what suddenly generates 3k Elandar Freezes.

I agree that 30% more respen would be a huge deal, but those are going to scale much smaller than that. The difference between Adept and non-Adept Wildfire at talent level 5 is 5% more respen for going from 6.5 eft to 8.0, the difference between Madness 6.5 and 8.0 is 2% proc. They matter, but they still usually scale poorly at high levels.

I don’t know yet what is causing the trouble. I’ve been hit for gobsmacking damage by rogue rares with Anorithil skills and no talents above level 5, but if I’m level 18 and they have Blood Moon, Lethality, and Shadowstrike (and are able to stealth consistently, unlike players) AND spell damage has already been improved then that could be the problem. But it’s not just that specific combo, either, and neither is it always with sky high talent levels. It feels like something more fundamental to spell damage has changed.

Yes to liking more tactical AI. But talents that were balanced around non-tactical AI (jelly, monolith, etc) will need additional love if they’re going to fit the same role in their class kit now that the ai just walks around them.

[Arcvasti]

High freeze damage is because freeze has a ridiculous formula that doesn't obey the standard diminishing returns for talent damage. And, on Elandar in particular, he now basically always gets Uttercold for extremely high cold respen alongside the bonus from the Awakened Staff. Freeze is getting mercilessly pummelled into reasonable numbers next patch, thankfully.

[anonymous000]

Side track a little bit: I just had a very short run with the 1.6.1 beta playing a wyrmic in Nightmare difficulty. It appears melee damage is now better, but accuracy still seem to be a problem. Even with 5/5 in Combat Accuracy I still missed quite often

[Snarvid]

Mana Thrust 1-shot by Pale Drake at level 39 as AM summoner (on the theory that hiding from enemies behind minions might help), 1263 damage (just 12 over my max hp). Resolve appears not to be helping against 1HKOs.

Unable to do enough damage to kill Massok despite snuffing all his Unstoppables with Healing Nexus, keeping him fairly steadily under Confusion effects and Insidious Poison, debuffing his physical and fire resist, and packing (among other things) talent level 9.8 Fire Drake with 308 Strength.

I do not recall having such trouble with Backup Guardians in 1.5.