Enchants generation general discussion

[zebramatt]

I posted the following on Mantis but I'm aware it's very much a nascent idea which is better off being discussed on the forum (assuming there's any interest at all!).

0006209: Enchants - implement player options

The following strikes me as non-trivial to get polished and working right so it may be a post-v1.0 feature. Or, if there's enough interest, maybe not. Here we go!

Presently whenever you pick up an enchant container (of a certain sort) you get points towards an enchant. Once a certain set of rules are met, you are given something shiny to equip somewhere on your character.

This idea is that once a (probably slightly modified) set of rules are met, the game pops up a window and present you with two (maybe more?) enchants, of which you may choose one.

The main guts of the suggestion, however, is the way enchants are generated and suggested to the player. The idea would be that any time the player is offered a choice between enchants, each of the suggestions is varied from and weighted against the other(s) - presenting the player with a very real choice every time.

In addition, I'm of a mind that stashes should no longer contain enchants - and there's just one way to get new ones, by collecting enchant containers and making choices in the direction of the stash enchants, if those are what you want. There could be some logic in the enchant generator to ensure you had more propensity for stash enchants of sorts you didn't currently possess, or something.

[x4000]

The stash enchants are not going away -- those are the "supplies" kind of enchants, and there are precious few of them as it is. For the rest... Player choice can be a good thing, so this could be a very interesting idea.

[zebramatt]

The stash enchants are not going away -- those are the "supplies" kind of enchants, and there are precious few of them as it is. For the rest... Player choice can be a good thing, so this could be a very interesting idea.

I assumed that's what you'd say, which is why I didn't include it in the mantis suggestion in fact! 

[Terraziel]

I'm behind this, it sort of makes enchants feel like a part of character development, and will cut down on the number of useless enchants a player acquires.....

Maybe add a button to not take an enchant this time, and let X% of the points carry over, I'm thinking of the situations where none of the enchants shown are of value to you, say because they boost elements you don't use.

[x4000]

Keith and I have also been talking about this internally -- we know we need to make some changes to the enchant containers -- but have not come to a resolution yet.  And in general have been too busy with other stuff to get to it in the past week, but it's one of a handful of mechanics things that we want to really get to ASAP in beta phase 3 (that, the mysteries, the more unique external generation, and monsters not walking off certain cliffs are the main ones that come to mind).

Some of my thoughts:

I want it to feel somewhere more Diablo-ish than I think it does now, but I don't know how to go about that.  With borderlands it was nice because I was constantly finding new stuff, to the point that I was comparing a couple of guns and then dropping the ones that I liked less.  That sort of experience would be nice to have here, versus so much being out of reach.

To that end, having way more modifiers possible on each particular enchant would be great.  Different gradations of power bonuses, etc.  I'd love it if we could have thousands or hundreds of thousands of possible combinations of enchants by the time we hit 1.0, and I don't think that's that hard.  If you have a lot more subtle variants of power boosts, for instance (1% gain all the way up to 100% gain), then that's 100 extra possible modifiers right there.  If you apply that to 20 different base enchants alone, that's 2000 unique enchants there.  If you're having more powerful (tier 2) enchants that combine two effects rather than one, but never overlapping effects, then you're getting to having 200k unique enchants right there.

I'm not saying I definitely want a gradation of 1% for each type of power boost, but that sort of general idea where there is truly unique-feeling loot (and where players could NEVER collect every variation of everything, no matter how long they play) is my general goal.  That's how Borderlands felt to me, and when it comes to a procedural loot system that's how I think of it.  And I think your system is already almost there, aside from two problems:

1. Not enough possible combinations yet.
2. After a certain point it gets extra stingy.  I know I asked for this, but I think that was a bad idea.  We need the early game players to get a bonus to how fast they get enchants (so that they have a certain baseline pretty soon in -- so maybe an enchant every 15 minutes or so when that player has earned fewer than 20 enchants?  Literally adjusting the costs on a per-player basis?), but then just sticking with the 30 minutes target baseline from there on out afterward.


That's more or less the experience I'm looking for: hook them with lots of rewards at the start, then give them a steady flow of rewards after that honeymoon period, and never diminish below that second amount.  Make it so that they're filling up their enchants inventory and tossing crap out because they have so much of it.  And make it so that by the time they hit tier 3 enchants or whatever (with three optional modifiers, perhaps), we're getting into the millions or billions of potential unique enchants territory.

And of course the idea is to take a few basic modifiers and then just combine the heck out of them, naturally -- so these things are getting so incredibly combinatorial in their nature simply because of the fact that they are getting multiple boosts per enchant.  If you did power increases in increments of 5% instead of 1%, that's only 20 different enchants, and we still hit the hundreds of thousands mark on tier 3 enchants just from that alone (20x20x20x20 = 160k).

Further thoughts:

However, in testing the current system with getting my double and triple jump enchant-items, I was struck by something: they were filled up with all sorts of extra conditions by the time I unlocked them.

These seems really undesirable in terms of flow to me.  The whole idea of this "spending budget" seems a bit incomplete.  Parts of it make perfect sense, but at the same time, you wind up with situations where you get lame enchant-items of one kind, and then awesome enchant-items of another kind.   Sometimes I just want a plain double-jump without a bunch of added baggage.  And then the added baggage is interesting to find once I've found the plan double-jump.

So what I'm trying to say is, whatever my first encounter is as a player with a new-to-me kind of enchant-item, I should be getting the most basic version of it.  Regardless of whatever is going on with the rest of the game.  If I'm on continent 11 and I've got mega-awesome enchant-items of other types, and then the game gets a new triple jump enchant-item coded into it, I should be getting the basic triple jump only.  And then next time it decides to give me a triple jump enchant-item, it should come with a larger budget and more enchant-effects.  And so on.

So in other words, while I like your idea of the budgets in a lot of respects -- and that seems to be a great way to ensure that power of specific enchant-items grows in a satsifying way -- that doesn't work across enchant-items in a way that strikes me as pleasing.  It jumps too many steps, and also has a tendency to make it so that the lower steps of many of them are never seen at all.

Even more to the point, I'd like to have the game have a prejudice toward giving me lower-budget enchant-items prior to giving me higher-budget enchant-items.  All of this argues for potentially a more complex system of remembering enchant budgets.

So here's what I'm thinking, based on the above:

1. For each enchant-item, we need to know on a per-player basis how many of that enchant-item they have collected at each enchant-item mark.  So if mark I enchant items have fewer effects and so on, then great.  The system you proposed for this would still work just fine in my model.

2. Once you have collected some number of enchant-items at the current mark for that enchant item, it would move you to the next mark.  So once I've had 5 different mark 1 double-jump enchant items out of enchant containers, then it starts giving me mark 2 double-jump enchant items.  Once I've had 10 of those (or whatever), it moves to giving me mark 3 double-jump enchant items.  And so on.  Individually for each kind of enchant item.

3. When the game is deciding what kind of enchant-item to roll for me, it just picks randomly from all the kind of enchant items (of course, honoring any rules caused by things like Fire Seeker or whatever).  There is no concept of budgets here, or overall.

4. Now the game looks at this enchant-item that it has randomly selected for me.  This enchant item would have a budget associated with it based on its mark level, I presume; or however you want that to work.  But at any rate, now it does it's per-this-specific-enchant-item logic and sees if it can come up with an enchant-item to work toward that isn't a duplicate of what is in my inventory at the time, and which meets the rest of your criteria based on budgets and current mark level for that enchant item, etc.

5. If #4 successfully found an enchant item design to use, great -- continue on as you had planned.  If it failed, then remove that kind of enchant-item from your working pool of potential enchant items (for purposes of #3 above), and then go back to step #3.  Keep repeating steps 3-5 until the working pool of potential enchant items is exhausted (incredibly unlikely), or until a viable enchant is found.

Thoughts?  I'm looking for a gradual improvement in enchant-item quality per enchant-item, is the bottom line, rather than it being something that is global for my character.  If I've been fire-seeking for a long time, and switch to Earth-seeking, I should have to start at the bottom.  And every time a new enchant-item type is added, all players should have to start at the bottom.

In general, all of the above is still valid (or not) regardless of whatever we do with presenting specific options to the players at the time they receive. An enchant.  Further thoughts in a second.

[x4000]

Having two, possibly three options when you receive enough enchant points is an interesting idea.  What happens, then, though?

1. Does the player have to pick one of those options?

2. Does the player get to say "pick one of these options or keep my points and defer until next round" like in Zelda II's upgrade system?

The first option is interesting in that it gives a modicum of choice, but it's still a very limited amount of choice.

The second option is a lot more complex, because we don't want players to be able to just defer 5 times and suddenly have some incredi-awesome enchant from the start of the game that gives them a very distinct advantage.  That was a cool thing about the Zelda II system, in that I think it didn't let you get more than one tier ahead in a given category before you had to upgrade the other categories.  IIRC.  But with the whole concept of enchant value, it's not fitting into neat tiers, and instead is a much looser mathematical concept.

And I'm not yet sure quite what to do about that, nor do I have time today or tomorrow to really discuss it or think about it any more than I just did, but it's an interesting concept and something I'd like to see working in a really compelling way.  So those are my notes for the moment.

[Terraziel]

If we take Diablo 2 as a model, then each act essentially restarted the loot quality, at the start of act 1 you get white leather armour and you move up through the colours to rare and unique leather at the end of the act, then at the start of act 2 it restarts with the new item types, so i am getting white chainmail and then up through the colours again.

Now obviously thats a pretty simplified interpretation of the loot in diablo 2, but it mostly holds.

What you would need to do is define the "acts", so maybe double and triple jump don't even start spawning till continents 2 and 3, but on a more basic level, maybe you can only get 1% - 15% upgrades on the first continent, with it switching to 10% - 20% on the second and so on.

[zebramatt]

It sounds like there's an argument for gating enchant types, I agree.

[tigersfan]

While I like the idea of gating enchants, IMO, there aren't enough enchants in the game to do that with... yet.

[x4000]

Also true.

By the by, "tiers" of enchants will be gated.  In other words, you have enchant-items and enchant-effects.  One enchant-item can have 1-n enchant-effects applied to it.  Higher-tier enchant-items will have more effects, making them more valuable.  That's how we planned to do a progression with the second and third continents and so on.

[Terraziel]

How many enchants do we actually have at the moment?

for comparison D2 has 100+ prefixes (That's essentially all of them bar +Skill ones), and about 100 suffixes (though those are a bit hit and miss) with close enough correlations to AVWW mechanics to just be lifted wholesale, most of them are quality changes but there are still 30+ distinct types.

In fact, as I investigated them It's quite interesting to see how they have done all their variations, essentially various calculations (based on a veritable slew of variables) limit the affixes and the affixes limit the range, but don't actually assign a distinct amount, meaning two identically named items can be different still.

Given quite how much D2 I have played over the years it's the sort of thing i should probably already know or at least be more cognizant of ....

[x4000]

We have something like 12ish base types of enchants, and something like 15...ish modifiers at the moment.  But both will grow substantially in the coming weeks.

[Martyn van Buren]

In terms of player choice, how about limiting enchant inventories?  I know we talked about this and ultimately rejected it for spells, but enchants could be a good place for it --- I imagine you'd have enough inventory slots to carry two enchants for each use-slot and a few special inventory slots for the vital utility enchants.  This would create a simple but difficult player choice --- do I take this or keep what I have.

[x4000]

In terms of player choice, how about limiting enchant inventories?  I know we talked about this and ultimately rejected it for spells, but enchants could be a good place for it --- I imagine you'd have enough inventory slots to carry two enchants for each use-slot and a few special inventory slots for the vital utility enchants.  This would create a simple but difficult player choice --- do I take this or keep what I have.

Definitely no, and I'm not going to touch that with a ten foot pole.  What we have are limited enchant slots, so you have to choose between what you can use at any given time.  But limiting the inventory itself just creates backtracking and lots of pointless frustration.  It's punitive to the player for no particular good reason, whereas the enchant slots cause the same sort of choice you're looking for but without making players keep a stash of enchants they might want later in some far off place.

[Terraziel]

I think the further point would be that if you are doing your item generation right then players won't want to keep stuff hanging around anyway.