Relative worth of stuff

[fizzix]

I didn't really want to muck up the 'what should we revert' thread with a discussion of this, but I really wanted to discuss this topic. Mainly I want to discuss Magnate's post:

All prices are based on "equivalent damage output". More on this later.

All weapon damage modifiers are priced according to the amount of additional damage they generate across the monster population, adjusted by relative monster power. So slay troll < slay orc < slay giant < slay dragon < slay evil, as you would expect. Same for brands, extra blows, etc. etc. Recently combinations of slays have received an extra premium, because in pure mathematical terms each slay is *less* useful than the last, but we all know that in game turns the player's intelligence in choosing what to fight makes a big difference.

So yes, the same slay on a large weapon is worth more than on a small one.

Resists are currently valued arbitrarily on a scale debated heavily on rgra in 1999-2001. Poison and confusion are at the top, and yes the low resists are worth less than the high ones. I am not intending to reopen that debate here, because soon I will be adding the empirical calculation of the value of resists along the same lines as the calculation of the value of the slays, so the debate will be moot. No it won't be perfect, and will need adjustment, but it'll be starting from an empirical standpoint rather than an arbitrary one.

Telepathy is currently valued at 35 damage per blow, i.e. you'd give up a ring of damage +15 AND you'd reduce your +25 weapon to +5, in order to have telepathy. To me that seems way more than it's worth, and here we have to separate the issue of relative power from pricing. You (one) would blanch at the price of a +35 weapon, yet still think that telepathy is underpriced. Same with boots of speed - they are enormously desirable, yet consistently underpriced. (Would you give up 30 damage per blow for +5 speed?)

So the approach isn't perfect - it's particularly difficult to conceptualise defensive qualities (like AC, resists, hold life etc.) in terms of "damage per blow". There are hacks needed to make the formula-based approach work in "real life" (by which I mean in the game) - as well as telepathy and speed, other examples are torches of Brightness and off-weapon blows/slays/shots. Those last are way more valuable than the mathematical addition to potential damage would indicate, and that's why 3.1.x has seen some overpowered randarts (the randart generator uses the same power algorithm - in fact that's where the pricing came from). I'm currently working on adjusting the power ratings of off-weapon damage boosts to reflect their extreme usefulness and prevent them occurring too often.

So the question is how to get values of certain abilities in terms of 'damage per blow.' A bit of a bizarre measuring stick, but it's the best we have.

I'm going to make an assumption. The assumption is that the cost should be based off of what a character who needs it the most would pay.

Speed: Magnate's estimate is +5 speed = +30 damage.

I'm going to assume that +10 speed = +60 damage. Assuming that a mid-game character is going to be the one to be most desperate for speed. This character is likely to be doing something like 50 damage per blow. +10 speed essentially doubles the damage output, so from a raw damage standpoint, +10 speed is probably around 50 damage per blow. Less than Magnate's estimate. However, +10 speed has numerous more advantages than just damage output. It allows you fight monsters you would be afraid of otherwise. It allows you to avoid getting double moved. It makes you 10 times more likely to survive teleporting into a room of gravity hounds. It makes you much more likely to sneak by a sleeping monster. And many more. For these reasons, I'd actually double the value of speed and set it at least 100 damage per blow for +10 speed.

Extra shots: Magnate's estimate (unknown)

This is easy, assuming that you are doing around 250 damage per missile, extra shots doubles that output. Assuming a 5 blow character, this gives 50 damage per blow for extra shots. Since ranged attack damage should probably we worth more than melee, extra shots should be bumped up a little to 75 dpb.

Extra might: Magnate's estimate (unknown)

Doing the same as extra shots, extra might will add about 60 damage. Dividing that by 5 gives us 12 dpb. Giving it the missile bonus, and we get somewhere around 20 dpb.

Telepathy: Magnate's estimate: 35 dpb

If the choice for me was how much damage would I need to be adding from a =dam for me to wear that over a fictional =telepathy. I really have trouble answering this. I might start considering it at +50 to_dam. Maybe +35, but probably not. However, if you instead ask, if I was at 0 speed, how much of a speed ring would I trade telepathy for, the answer is probably somewhere like +7 - +8. Which means, if you buy any of my speed analysis, telepathy should actually be valued something like +75 dpb.

Magnate says he's reworking resistances so I won't try to mess with these.

[Pete Mack]

With things like ESP vs damage, it's not either-or; I would definitely carry a swap if I could get an additional +35 damage/blow by giving up ESP. I would definitely not carry a swap if I could get +35 damage/blow by giving up +5 speed (unless I had speed > +15 to begin with.)

You are trying to measure things that are incommensurate. There's just no way to do it.

[fizzix]

With things like ESP vs damage, it's not either-or; I would definitely carry a swap if I could get an additional +35 damage/blow by giving up ESP. I would definitely not carry a swap if I could get +35 damage/blow by giving up +5 speed (unless I had speed > +15 to begin with.)

You are trying to measure things that are incommensurate. There's just no way to do it.

I agree, but Magnate says that everything gets measured in damage per blow. So that's the metric I used. When it comes down to cost, everything gets paid with the same unit, so we need to figure out a cost base for each item. My point, if I have one, is that speed and ESP are more valuable than they are currently estimated.

The other thing I was thinking of is a log base scale instead of a linear one. I dunno what the base should be though, nothing really put out such great numbers.

[Jungle_Boy]

This is a very tricky issue. The amount of damage you are willing to give up to get a certain attribute is heavily dependant on how much damage you have. If i'm doing 35 dpb would I give up 35 to get ESP? No. In my experience what I'm willing to give up to get something new, whether it's damage or resisitance or ability is whatever I have a bunch/too much of.

I propose a new selling pricing structure, 1 AU per ability, resistance, AC, avg damage, charge

[fizzix]

This is a very tricky issue. The amount of damage you are willing to give up to get a certain attribute is heavily dependant on how much damage you have. If i'm doing 35 dpb would I give up 35 to get ESP? No. In my experience what I'm willing to give up to get something new, whether it's damage or resisitance or ability is whatever I have a bunch/too much of.

I propose a new selling pricing structure, 1 AU per ability, resistance, AC, avg damage, charge

That's why you can't look at it as 'giving up damage' but instead wearing a damage ring vs. wearing a ring that gives this extra ability. It still has problems, but not as many.

[Tiburon Silverflame]

The root problem IMO is trying to fit everything into a single equivalence class, especially across all item types. And as Pete and J_B are suggesting, the value of any ability depends on what else you have.

So, it seems like the better approach is to consider a 'typical' near-end-game equipment set. Is ESP on a weapon, worth +20 damage in this construction? No, because in most cases you'll have ESP from something else. Also, damage is *mostly* determined by the weapon, with the exception of off-weapon extra swings. Yes, Hammerhand + Fingolfin means +19 damage, and sure, that's nothing to sneeze at...but in an end-game kit, this is probably only about 20%, from *2* slots. The weapon's still 80%. With an off-weapon brand, the tricky part is assessing how often it will apply. If I've got a weapon with 2 or 3 slays and brands, the off-weapon brand is frequently redundant. (This also points out why an off-weapon extra blow is so strong. Worst case, it's +16% from one slot, and it *always* applies.)

Similarly, while +5 speed on a weapon is wonderful through the mid-game, by somewhere around DL 60 or so, it's losing luster. However, we'll still consider a +10 speed weapon for an end-game kit, even if the damage is marginal in itself, because it will improve flexibility in other slots.

(Oh, a huge note here: as long as archery is so overpowered, +10 speed on anything but your shooter, is highly valuable.)

[RogerN]

I like the new pricing system. I also think the old prices were OK (with the exception of ridiculously expensive =Speed). Prices are always going to be arbitrary; as others have already pointed out, you can't really quantify "usefulness" as it depends too much on context.

In fact, I think any arbitrary pricing system will work, provided that prices fall into a reasonable range of affordability which is appropriate for an item's depth and/or rarity. I think that the range of "reasonable" is pretty wide, too. The price of an object is just another variable which factors into how useful it is to you in your present situation.

[fizzix]

I'm not by any means proposing an overhaul of the pricing system. It's not at all broken. Instead I'm just proposing some introspection into how prices are set, leaving the door open for a couple tweaks.

[Magnate]

I'm not by any means proposing an overhaul of the pricing system. It's not at all broken. Instead I'm just proposing some introspection into how prices are set, leaving the door open for a couple tweaks.

If you want to do this properly, please read http://trac.rephial.org/browser/trun...ct/obj-power.c. You don't need to know how to code in C, just how to follow IF/ELSE logic. (You can skip the slay evaluation and start at line 229 after looking at the #defines.)

To answer Tiburon again, yes, the assumption has always been for 'near endgame' kit - that's why we assume five blows per round for melee. I don't mean to be rude, but I haven't thought about this quite as little as you think I have.

[Tiburon Silverflame]

And you're missing the major point we're all making. The value of an ability is HIGHLY dependent on context. The whole point of a near-endgame kit is to make assumptions that will usually be true. THAT means, I've got ESP from a ring, amulet, or helm...so ESP *on my weapon* is of VERY low value.

It is NOT true that an ability is always worth the same amount, regardless of item type. It is also true that an ability's value is *not* constant, again with a high correlation to the slot. That is: see invis is a nice property for an early-game weapon...but becomes valueless late. So, on a weapon, if the weapons value *as a weapon* (that is, considering damage, blows, and speed ONLY) then see invis and many other abilities, carry full value. But if it's high, then they add much less.

So proper valuation has to take a lot more into account. And a lot of this is there, at least in some fractional form. You already have a sense of classifying the power boost into low resists, high resists, and miscellaneous within the use of ADD_POWER2, and damage and (I assume, can't see it just at the moment) speed have their own evaluation sections. Great. That's useful.

What I'm suggesting is, *rather than combining them all the way through*...you track them separately. So you get ratings for damage, stat bonuses, defenses (resists/immunities/sustains), miscellaneous, and speed. Final value is derived as a combination of everything, *but* the weights applied to each factor are slot-dependent:

Weapon: damage 100%, stats 40%, defenses 30%, misc 20%, speed 50%
Armor: damage 20%, stats 50%, defenses 100%, misc 80%, speed 30%
Hands: damage 50%, stats 50%, defenses 50%, misc 50%, speed 50%

And so on.

[Antoine]

And you're missing the major point we're all making. The value of an ability is HIGHLY dependent on context. The whole point of a near-endgame kit is to make assumptions that will usually be true. THAT means, I've got ESP from a ring, amulet, or helm...so ESP *on my weapon* is of VERY low value.

Tiburon please don't "shout" at Magnate. Also I don't agree that we are "all" making this point...

A.

[Derakon]

And you're missing the major point we're all making. The value of an ability is HIGHLY dependent on context. The whole point of a near-endgame kit is to make assumptions that will usually be true. THAT means, I've got ESP from a ring, amulet, or helm...so ESP *on my weapon* is of VERY low value.

Awesome. So when the weapon shop happens to generate a Weapon of Gondolin with ESP, I can get it on the cheap.

IMO, what we should do is simply tweak the prices and the availability of cash until we can look at a few standout items (caps with telepathy, holy avengers, gondolins, boots of speed) and say "Yes, that costs about the right amount." That's good enough.

And Tiburon, I don't think Magnate's missing your point so much as saying "Yes, I've already spent a lot of time thinking about this." Just because the current iteration is partially flawed doesn't mean he hasn't taken into consideration a lot of things while making that iteration.

[Jungle_Boy]

Since items are of differing value throughout the game and things like see invis and FA are awesome early but of almost no value later since you are almost guaranteed to have it somewhere does that mean we should add a modifier for dlvl to the pricing algorithm? As you go deeper things get cheaper Talk about encouraging diving!!

[Atarlost]

There's no reason not to price stuff based on what dungeon level would have the highest price. Money shouldn't be an issue to deep characters anyways.

This would require using some different values than the randart generator though.

[Timo Pietilä]

I don't see any great relevance in this whole issue. This game is not about shopping OTOH old angband had discounts and I miss them.

Value-based or fixed price / ability, who cares? There are no huge errors in prices in 3.1.2v2. Looks like there is a heated discussion going on on non-issue.