Implications of Upcoming Farming Nerf

[SageAcrin]

First, a general statement/caveat that sums up the main thread of this discussion: No, I don't think the current situation is ideal with item dependence. A lot of this can be fixed with some fairly minor changes to artifacts, which I'm currently collaborating with Stition/Waladi/Shibari to look at (if anyone else wants to poke in before we finish and make a post, I can give you access). This post was primarily to point out the item-dependant nature and what these farming changes would mean, and just to point out that it's a fix that while understandable (no one likes farming), raises some other issues.

Some of it is more basic class/mechanic issues which are harder to deal with. You're right in that there would be no easy, sweeping changes that would fix it, nor is it necessarily good or desirable to do so due to the critical nature of variety. I have a lot of ideas on this topic, and I think there are a bunch of ways that you could do most of it without radically effecting the balance on normal difficulty, but I'll go into them in another post.

Fair enough. I'm (obviously, given some of my past topics! ) not adverse to buffing artifacts or egos, particularly in ways that would enable a little more variety/reliability on Nightmare and Insane.

Also, some skills that aren't that notable on Normal are moreso on Nightmare/Insane; Usually not knowing an enemy there is merely distracting or damaging on Normal, not fatal. On Nightmare this is much more dangerous and can only get worse on Insane, I'd imagine. Such info skills are, as such, fairly safe to put on artifacts. Mobility skills are a little weirder(as they're good on all difficulties), but there are certainly some artifacts that are "low end" functional that could be given such skills.

If you can come up with good ideas on skills, though, I'd really love to hear that. That's a lot harder.

Oh, and only the last boss can crit for ~5k, regular crymancers are in the much more reasonable region of 2k.

Hey, that sounds familiar. They hit for 1200~ on Nightmare with crits(and that was the very top end, 1000 on a crit was more common. This was with fairly good 50% range cold resistance.), whereas Elandar crit for 1700 through 70% cold resistance.

(Unlike most everything else later in the game, cold resistance actually sorta works on Elandar, because he "only" has 30% resistance penetration. Whereas even on Nightmare 40% is the norm. The difference is oddly notable.)

Lastly, as to the whole perfect play issue, I think we're getting caught up in semantics again. Bottom line: Playing very well/strategically/intelligently throughout the game should net you some good win percent. It should require exponentially better play to require the same win percent as you go up difficulties, but in general the amount of randomness should be poker-style instead of getting-hit-by-lightning style

Yeah, I can agree with that. It's a lot harder to achieve than talk about, but optimally, I definitely agree with that.

[Stition]

I've played so much tome, over 90 days, and that's not including time spent offline I think. Shibari got me streaming my games, and since I know it's terrible to watch a stream of a turnbased game (other than to watch me die), I started making some more vods to help people get over the hurdles I had to face when learning insane.

I was pretty discouraged when insane was nerfed originally since I was so close to being able to win (got a berserker to the last event, but died). Even though I have no time invested in the development of the game, I still feel like "it's my baby." Other people probably feel similarly since it's such a well designed game.

If people are interested I put up 2 more vods (they'll increase in content quality when I get more comfortable with everything):

Pet abuse in Gorbat Bastion:
http://www.youtube.com/watch?v=2gUxRMPFuo4

Using beckon to position enemies in Vor Bastion:
http://www.youtube.com/watch?v=OxvCfYMc4XI

There isn't much activity in the spoiler forum, so was hoping things like this would help people. I'll add actual commentary explaining the logic behind certain decisions when I get more time and the community completes the insane Darkgod challenge.

[jenx]

I find this conversation very interesting. I started playing Nightmare about 2 months ago, having cleared about 10 times on normal (tw, slinger, solipsist [my favorite], brawler, cursed, mindslayer).

I have really enjoyed the challenge, and it has improved my playing immensely. For example, when I wrote my dwarven solipsist guide recently, I ran through 3 dwarf *normal* roguelike solipsists. The first won, the second died stupidly at lvl 19, the third won. Now, I *never* could do that beforehand. I feel that on normal roguelike, with a solipsist, I could win comfortably and regularly now.

So nightmare has this very important role to play.

I've never played insane though, cause I still can't unlock it. I should try nightmare adventure mode, but I hate missing out on the achievements.

In my view, the range of classes and races now is excellent, well-balanced, and provides consistent enjoyment to huge numbers of people. And I think development should always focus on this main game. There is too much at stake to make fundamental balance changes at this level (remember when rares first appeared? I stopped playing THAT version.)

I actually think Nightmare is very well balanced now - it provides a MUCH greater degree of difficulty than normal, but is winnable (though I have yet to win!). But I can see it is within reach. And like the others who play this, my view is that the number one problem is massive burst damage. And it is not just 5000 and beyond. It is 250 on level 1, etc.

I think one of the BIG problems in nightmare upwards is that resists mean so little, because so many things either hit for si much it doesn't matter, or they penetrate. I think this should be addressed, as it makes the whole exercise of resists somewhat pointless.

I should also point out that I don't use item vault, and this seems to be a fairly key factor for the wins on insane, as I've looked at them and they use lots of item vault items. I don't even know how this works, tbh, but it doesn't quite seem right to me to win in a rng game, but item vault means using randomness from across multiple games in one game.

[grmblfzzz]

@Stition: I'm amused by the little popup commentary on the video's. Definitely helps a lot, and explains some of happenings that would otherwise be hard to catch due to speed/resolution.

@Jenx: You pretty much touched on several of the core issues. First, I'd like to say that I agree completely on the nature of how higher difficulties work and how valuable they are. It's pretty impressive if you really focus and struggle clearing something that is challenging for you, when you take a step back down to see just how much of an improvement was made. That's why I'm in general just all for an infinite slope of select-able difficulties!

For overall balance, I agree, the game is in a fairly good place right now. Of course, that's not the same as saying that it's in an ideal place, as there's always room for improvement, and good balance while maintaining variable playstyle. That said, there are still some fairly big gaps, which just become more apparent as you move from Normal -> Nightmare -> Insane. I totally agree that things shouldn't be adjusted for the highest that might negatively effect the lowest, which means generally that sweeping changes are out and you have to use more case-by-case subtlety. Probably tomorrow I'll make another post about some of the class balance issue's, and we can go into it then.

The basic issue you brought up with burst damage is more or less is what is being addressed. I tend to use end-game examples because that's what consumes the most thought for me, but you're right that it applies earlier as well. For instance, Wrathroots can currently crit you for ~300-350, depending on resists. It's real hard for most characters to have that much life by the time of Old Forrest. Increased difficulty like this is part of what I like, but it again brings up item-dependancy and class balance issues. Some classes can shrug that off, but others really have to scrounge for some cold-res/Hp-boosts/warding wands to safely do stuff like that.

So... to wrap up the point, the game IS currently fairly balanced on Insane... BUT, it is in such a way that is fairly item dependent. Addressing this is pretty complex, but there are steps being taken. There's a general ego buff going in, and we're talking about taking a look at a lot of the legacy/under-powered artifacts and adjusting them so it's less completely rng based.

Lastly, just a quick word on the item vault: If you donate, there's a nifty little orb in your Sher-tul that will let you transfer items between chars. Mind you, they have to get to shertul, and they still get lost if the char dies with them. Overall I think this is an amazing thing, but for convenience rather than sheer power. If you're a madman and cared to farm (See whole point!) as of this current version, all artifacts tier 2+ could be guaranteed with enough farming, which while undesirable was somewhat necessary. Also, I'd like to point out that it really only helped my and Stitions 2nd+ winners! My first dude had to spend probably ten hours farming for items I deemed critical, and I think he had to farm for more. So the vault just compresses the amount of tedious farming we had to do.

I'm all for super manly, no vault wins. Just the whole point was that since the items are a little TOO random in their power at the moment, I for instance almost certainly COULDN'T have won on my summoner just cause he got unlucky in his playthrough before I had to start farming. See my posts with sage-acrin: Poker-style randomness good, getting-hit-by-lightning randomness bad!

[Parcae2]

Hi! As the troublemaker who requested the grinding change, I thought I should weigh in. I'm an Insane player too, albeit not a very good one - my best character to date died at level 32.

This discussion is exactly what I was hoping would happen when I made that request. If grinding is possible, then the game becomes balanced around grinding. If you take out grinding, then it forces the game devs to rebalance the game around the new normal.

Anyway, on the more specific question of Insane-difficulty balance: In a certain paradoxical way, I would argue that Insane is actually more balanced than Normal from some perspectives, and here's why.

On Normal difficulty, 99% of encounters are easy for a competent player. This is problematic for two reasons. Firstly, it's boring. Secondly, it lures the player into a state of complacency that often (in my case, at least) results in their not realizing that they are in one of the few actually problematic encounters until it's too late. Most roguelikes have this problem, which is why they tend to be quite short (if you don't die). Constant concentration is difficult; constant concentration when it is only rarely needed is impossible for most people.

Autoexplore reduces the problem by removing areas that are empty from consideration, but it doesn't eliminate the problem, and it also forces the player into suboptimal play (because scouting with Arcane Eye or Track is better than autoexploring) for reasons of convenience. This is a problem, and it is also, I believe, why grinding was removed in the first place - players should never feel that they have to do something boring to win.

In my opinion, the solution to balancing Insane is the same as the solution to balancing any difficulty. Every encounter should be very difficult; none should be impossible. The most fun I've had on Insane (or in the game, period) came from going through Dark Crypt a few days ago on an Oozemancer, and I think it was a microcosm of what the game should be like on that difficulty (or, accounting for player skill and experience levels, any difficulty). I was never able to relax. Every kill felt rewarding. Yet, at the same time, I was never forced to resort to exploits or unusual tactics, and I had not had to grind to get to that point.

In order to achieve this, I think that:

1. The game should incorporate an algorithm to spread rares more evenly through zones. Insane difficulty should have even more rares than it currently does. Scouting around every corner only feels like a chore if it's pointless a lot of the time, which it currently is, even on Insane.
2. All classes should have improved access to scouting abilities.
3. Across-the-board item rebalancing (already underway - thanks, Purequestion). Some items should be guaranteed, e.g. Withering Orbs or Blood of Life.

Bottom line: higher across-the-board difficulty, less difficulty from fluke encounters (three rares at the entrance), less item randomness.

[PureQuestion]

(already underway - thanks, Purequestion)

Hehehe, SageAcrin, Darkgod, and all the people who contributed to that artifacts Google doc are just as deserving as thanks.

As for "already underway", it's hopefully mostly finished (barring a couple stragglers).

For now.

[grmblfzzz]

Parcae2 brings up some interesting points, which I want to (belatedly!) respond to: Fundamentally, I suspect we're in agreement.

Grinding is tedious and no one likes it, but when it existed the game was tenuously balanced due to it's universal availability. If you were capable of playing sufficiently tediously, you could get the tools on any character in order to succeed with him... But since, as mentioned, that is not fun, it's ideal to try to actually balance through normal means. This patch should do a good job of reducing item-based RNG by spreading the quality of artifacts better, and the same things with prodigies. More options and less RNG all around.

On the topic of difficulty spread/consistancy: I see what you're saying, and I agree that it would be ideal... but there's really no reasonable way to avoid difficulty spikes while keeping in any amount of interesting variations. You'd have to make the levels set, because the difficulty of encounters varies wildly with terrain. You'd have to control what can spawn where, as for instance starting in a big room chock full of mobs is infinitely harder than an empty corridor. The classes/enemy AI would have to be radically more balanced, as some are simply harder than others (and it further varies depending on what class you are). You'd have to, as you suggested, de-clump/spread rares around. These things aren't really either feasible or desirable on various levels.

However, what you can and should most certainly try to do, is make it so that the tools to deal with any situation are available in an ideally non-random way to characters. This is really the heart of this discussion. As mentioned, currently they are... if you are willing to grind. Overall the art/prodigy/scaling changes should reduce this problem significantly in the patch. But you bring up a specific issue I've been meaning to address: the availability/spread of scouting abilities. One of the best things about ToME is that you pretty much do have the tools to deal with any encounter (that's not you being a madman and diving out of depth!) if you can make some tactical decisions before shoving your face into a room, that is.

I'd like to break this into two parts, as there are really two separate issues related to this scouting issue.

Availability of Scouting

A majority of classes actually have some intrinsic scouting ability (will discuss the wide variance in quality later). However, there are a few that do not, and they are completely item or escort dependent. Scouting is what allows you to be tactical and make intelligent decisions, and thus as difficulty goes up it becomes exponentially more important. Anyway, these are the classes that don't start with any sort of scouting:

Arcane Blade**
Doomed
Reaver**
Anorithil**
Sun Paladin**
Mindslayer

** means that they're casters, thus no access to zigur, which means no access to track via escorts/worldly knowledge.

These classes are completely item-dependent, and have to at some point find either the arts that scout, or NightHunter/Delving/Arcane Eye. Conveniently, these items are pretty common so it isn't a huge deal as it'd be pretty hard to go through the game without finding one of them. It can definitely make the early game rough and aweful random, though, until you do! This almost-but-not-quite guaranteed scouting is close to not being an issue. I recommend changing Woldly Knowledge to be restriction-less so that you can guarantee getting the scouting of your choice if you so desire (I'm actually greedy, and always want *both* divination and track if possible, heh)!

20/26 classes that have some intrinsic scouting is actually pretty good, and much better than at various times in the past! I'm all for people coming up with interesting idea's to complete the process, though. The type of scouting available is really impressively varied, and I'm sure there are interesting unexplored options! Ideas: Please generate them!


Quality of Scouting

First off, varying quality of scouting is totally fine, similar to varying difficulty of classes. Also, different mechanics are just interesting and good, and lead to a variety of play-styles. For instance, you'll note that I don't have Alchemists and Solipsists on the above list. Alchemists have the ability to send their golemns around corners and such for them, solipsists can kind of do similar thing with their thought-forged, as well as poke through walls with projection. Good enough! Not as generally handy as track, but it fills the critical role of scouting without using your face Or take Arcane Eye, while it's not a handy AOE, it has compensating advantages of being instant and seeing into vaults. Again, totally acceptable.

However... I think some of the psuedo-tracks should be looked at. Specifically, many of them have a fixed (unscaling) radius which really, really hampers their use for scouting. The biggest offenders here is Meditation, and I also think Vimsense should be looked at. Reasons? On radius style scouting, there are many situations where sub-11 radius is worthless. That being all mobs sight radius, and there being many open areas/long corridors where such a radius actually doesn't allow for scouting.

Meditation: There are 4 wilder classes now, who depend on this as their natural scouting. It already has a disadvantage that you're doing half damage, so you can't sustain it while actually fighting most of the time... I honestly fail to see why the radius of it's track portion shouldn't be buffed. I don't mean necessarily up to an actual track-like level, but perhaps to a level corresponding to the radius of earths eye? Personally I feel that somewhere right around 14 (the radius of TW's precog track) is about the minimum level that it can really be useful and non-tedius. Minimum 11, so you can see mobs before they can see you, and that gives enough of a buffer so you don't have to do it constantly.

Vimsense: So, this has the advantage over track of reducing blight resistance, which is sweet. But, again... not actually usable for scouting in many situations due to it's 10 range. Further, it takes a turn, costs a whopping 25 vim, and aggro's stuff. I feel that the aggro aspect is the cool and unique part, and counteracts the blight pen. But as is it's really more of a combat buff than a scouting ability. I'd like to see the range increased to, again, at least somewhere around the TW level. Which would make it a kind of manly, self-hunted kind of scouting

[myndzi]

I haven't finished the whole thread yet, so please excuse me if I jumped the gun.

A word on randomness.

Some years back, I read an insightful post about music shuffling. (http://keyj.emphy.de/balanced-shuffle/) The gist of it is, when what we think we want is "randomness", what we actually want is somewhat more ordered, not more random. I wonder if the same concept cannot be applied here? The goal of randomness in this context is to add variety to the game. Ideally, I think, you would start a game with a given character, and complete it differently based on how the random spawns fell out.

This requires a couple things, though. One is that different builds are (somewhat equally) viable, to the extent that the determining factor is not "what can be viable" but "what equipment did I get/challenges do I need to overcome" you can "go with the flow" instead of struggling against it (farming). The other is that the items you randomly get work out equally viable. If you get a bunch of useless-to-your-class items or the wrong kinds of things, the option to go with what you get goes off the table.

What might be done about this is to have less randomness. The key insight is that less randomness doesn't need to equal less variety or less challenge. The question is, how to produce the right kinds of distributions?

Cataloging equipment by its utility, and classes by their equipment dependencies, is a good factor to take into consideration. You could, for example, create a more even distribution of equipment by kind while maintaining randomness of quality. The kind of the drops could be made more even, or perhaps even skewed a bit to avoid saturating classes with equipment that's useless in particular to them.

This process could restricted to populating stores, perhaps, as a means of providing access to the necessities without forfeiting the luck-of-the-draw of the actual dungeon crawling. Or [randomizable factors] could be categorized by other characteristics in order to better control the extremes of randomization.

There's an algorithm in use in Tetris called a history randomizer. It keeps track of the last four pieces generated and tries up to 6 times to NOT generate one of them. If it does anyway on the 6th time, it goes ahead and keeps that piece anyway. In practice, it's a pretty good way of maintaining unpredictability (can't rely on a specific piece coming up soon) while reducing what could be seen as 'unfair' sequences (long sequences of certain pieces or long absences). The end result is that the player can focus more on the actual task at hand - fitting pieces together - without a need to mitigate their play to provide for those extremes. The game is still plenty difficult, but it's the lessening of randomness that makes it enjoyable in a way that it couldn't be otherwise.

It seems like this idea could apply quite well to roguelikes, but since I'm not as familiar with the mechanics of this game as many of you, I can't suggest in detail how.

[b0rsuk]

I think Insane is called Insane for a reason.

[Doctornull]

A word on randomness. (...)

Great post. Welcome!

What might be done about this is to have less randomness. The key insight is that less randomness doesn't need to equal less variety or less challenge. The question is, how to produce the right kinds of distributions?

It's also possible to generate random sequences with different kinds of randomness.

For example, to generate a store inventory one might do the following:
- Read in a weighted list where A is very common and D is rare (A, A, A, A, B, B, B, C, C, D).
- Choose a random item from the list 8 times.

It's quite possible to end up with 8 instances of A, or zero instances of D.

Instead, one might do the following:
- Read in a weighted list as above.
- Randomly sort the list.
- Read the list until you see two different types, and add both of them.
- Repeat 4 times (for 8 items total).

Now you're guaranteed to only see A at most 4 times out of 8, though it is still far more likely to appear than D. You've reduced the variance by removing the independence requirement from the item selections.

It's possible to similarly reduce 'fat tail' events by variance reduction in other areas, and you're absolutely right that it's generally order shuffling which people want, and 'black swan' events tend to frustrate and annoy.