Randart luck :)

[Zikke]

Yeah the "blink, slam CLW, melee" technique after using up all of your Slay Evil or seeker ammo works pretty well if you're patient. My kobold ranger did it with crappy randarts and an ego long bow.

http://angband.oook.cz/ladder-show.php?id=10159

It's not a glamorous fight, but it's effective. Sometimes you may get only 1 or 2 rounds of arrows off on him before he's in melee range, but that's ok. The key is to minimize the teleporting so that he doesn't regen a lot of hp while running over to you (blinking is the way to go unless he absolutely fills up the area and you have no other choice).

[PowerDiver]

http://angband.oook.cz/ladder-show.php?id=10231

Heres a more typical endgame setup of my rather unlucky ranger. The artifacts are fine, defenses are covered nicely, but weapons are lacking.
- best shooter has been x4 with 20ish to damage, and 2 shots/round
- not a single stack of high base slay evil ammo
- 60ish damage per blow versus morgoth on melee weapons

Would you go for it at this point or keep looking for a better weapon/more healing potions ?

I've killed M with a ranger using only non-ego seeker ammo.

If you melee, or shoot and let him hit you, you don't need as much healing as you think. CLW is massively overpowered. ?phase and CLW repeatedly. You might need to ?phase twice to get your hp back up over 700. Or if down a lot say from mana storm CLW 4 time before using !heal. That sort of thing should significantly extend your healing potions. I'm not the only person to kill M with a mage using no potions of any type of healing at all.

The most important thing is to take a huge stack of ?phase. I suggest 99 if you can keep your speed high enough.

The best reason to keep searching for stuff is if there are any nasty uniques you have not killed yet.

[Estie]

http://angband.oook.cz/ladder-show.php?id=10231

Heres a more typical endgame setup of my rather unlucky ranger. The artifacts are fine, defenses are covered nicely, but weapons are lacking.
- best shooter has been x4 with 20ish to damage, and 2 shots/round
- not a single stack of high base slay evil ammo
- 60ish damage per blow versus morgoth on melee weapons

Would you go for it at this point or keep looking for a better weapon/more healing potions ?

[Magnate]

@fizzix: Tiburon is about right. The current evaluation of slays/brands is done like this:

1. Calculate the power rating of each monster - this isn't perfect but does take into account hp, damage output, frequency/packs, spells etc. (One day I will rewrite it but it will do for now.)

2. Express the total power of monsters vulnerable to the slay as a proportion of the total power of all monsters, and multiply by the slay mult to give an "effective multiplier". Hence slay troll has an effective mult of 1.013 while slay evil is about 1.6 - because there are many more powerful evil monsters.

3. Multiply the base weapon damage by that effective mult. Since object power is based on damage output, that reflects the slay in the object power rating. Tricky for off-weapon slays because we have to assume the weapon damage - this is currently just 3d5 but I have concluded that off-weapon slays/brands are actually far more valuable, so this will be changed.

So, I want to do a similar calculation for resists, using a factor which is the total power of monsters using that element as a proportion of the total power of all monsters. This needs weighting by the damage of the elemental attack and a few other things, but the most difficult issue is how to turn that number into a damage-equivalent for object power. This I think will have to be trial and error, and will have to address the rarity problem: the four base resists will numerically score much higher than the high resists, but they are so common that they are not worth as much.

[Tiburon Silverflame]

I would assume that the power evaluation means, assessing the relative value of the property. For a Slay, for example, consider the list of monsters for which that Slay applies.

--What is their overall power/threat? Or, it might be easier for an initial assessment to ask, at what point do the toughest members start to appear? The shallower that is, the weaker the slay.

--Frequency: how many appear, and how often?

--Other factors: what other Slays typically apply? What about brands? Vulnerabilities? What about special attacks?

Consider Slay Orc, Slay Troll, and Slay Giant.

Threat: The deepest orc entry looks to be Azog, at DL 23 if I'm reading the file here right. That's not very deep. Trolls start at DL 25, running through DL 40...mid-level tough. Giants go all the way to greater titans, and include 2 deep uniques.

Frequency: they're all fairly similar here.

Other factors: many orcs are vulnerable to light. A couple of the uniques have some resists, but otherwise I think they're light on special defenses. For trolls, a couple are vulnerable to light. There are some special defenses, and special attacks (water and ice trolls). Giants up the ante here that much more.

So it's fair to say that Slay Giant is stronger than Slay Troll, which is stronger than Slay Orc. Of course, that doesn't mean I always take a Slay Giant over a Slay Orc; I'll strongly prefer a Slay Orc weapon over a Slay Giant weapon, if found on DL 15...because I can use the Slay Orc, while the Slay Giant won't apply for so long that I can safely assume I'll find something better before I need it.

You can do something of the same assessment for resists, although it's less clear-cut. Here, it's a risk assessment: what is the risk for instant death, and what is the risk for a crippling side effect?

[Zikke]

For the record I just want to point out what was said earlier that the examples of randarts listed in these threads can't be considered representative of all randarts created, just those that really stand out which happen in a minority of games.

[fizzix]

* empirical power evaluation of resists (as is currently done for slays and brands)

can you explain what you mean by this for the slow of mind? You can assume a reasonable familiarity with the randart code.

[Magnate]

The only time I have found a PDSM was when I no longer needed it, which made me a bit sad.

But that's the nice thing about randarts, like you said the base item's change so you can get some nifty stuff much more commonly... which is of course tempered by the fact you can find a blade of chaos with rfire, some ac, and a fire brand.

Besides the fact that some scary powerful randart combinations can happen, so far I think Magnate has done a wonderful job, and that balancing multiple shots will make this all even less of a problem. (A set I made had a ring with +3 shots and there were two x6 longbows in the set as well.)

Thank you. I'm pleased with the way it's gone so far, though I acknowledge that as things stand randarts are slightly overpowered (I have created a ticket to address this). I agree with the person so doesn't like capping - neither do I, so I will use probabilities instead - +shots or +blows will be much rarer, and incredibly rare at pvals above 1.

That said, I am not going to do endless minor tweaking of probabilities or power ratings in the current randart generator. Having got it into the mainstream code and producing interesting randarts, the next things I want to do all involve major surgery:

* slot-specific power evaluation
* empirical power evaluation of resists (as is currently done for slays and brands)
* themed randarts (combinations of mods, some slot-specific)

The first will probably wait until archery rebalancing is done (which in turn probably means waiting for fractional blows). The others will get done whenever they can compete with my other priorities (re-doing pref files, packaging Dubtrain's sounds, etc.).

[ewert]

To be honest ... I'd vote for getting rid of all types of extra attacks anyway. It is too tricky to balance and too powerful for the wrong classes (ie. those that are supposed to be not as good in melee/ranged). I already suggested changing rangers from +shots to +might. Might there not be an idea of the same style for +hits too?

+might for melee weapons too? Acts like an extra "+1" slay multiplier? This would scale with base class fighting prowess so that warriors will still be that much better than mages at melee. It would not be worth very much with light weapons, but tbh I don't see much of a problem there, neither are slays either.

I really need to find the time to dig into Angband code and do my own "IVanilla". :P

[Zikke]

I agree with Estie that playing with Randarts is MUCH more fun when you don't look at a randart spoiler file and just see what you find. I would like to see a counter in-game so I know when I'm running out of possible randarts, but it is very exciting rushing to ID any randart just in case it's that one crazy overpowered gem.

[Estie]

This just is not true. When is the last time you found either Bladeturner or The One Ring? So how exactly does not randomizing these items affect gameplay in any way?

Its all about not knowing whats out there. Of course I dont find them in the vast majority of my games, but in the one I really do, I want to be surprised.
Winning ? Of course, thats neccessary; but I have done it before, so nothing new there. Some people start minimizing turn count at this point, others play ironman or competions; for me, it was the quest to find that perfect set of artifacts that has kept me going all these years. And if I didnt mind knowing in advance what it was going to be, I might as well play the standart game.

Imagine you found Ringil for the first time. Wouldnt you prefer it to not have read the artifact spoiler when you do ?

[Sirridan]

In all my years of playing Angband, I have never once found either Bladeturner or the One Ring.

I have found Bladeturner-based randarts twice however, because the new base item is much more common than PDSM.

The only time I have found a PDSM was when I no longer needed it, which made me a bit sad.

But that's the nice thing about randarts, like you said the base item's change so you can get some nifty stuff much more commonly... which is of course tempered by the fact you can find a blade of chaos with rfire, some ac, and a fire brand.

Besides the fact that some scary powerful randart combinations can happen, so far I think Magnate has done a wonderful job, and that balancing multiple shots will make this all even less of a problem. (A set I made had a ring with +3 shots and there were two x6 longbows in the set as well.)

[Zikke]

This just is not true. When is the last time you found either Bladeturner or The One Ring? So how exactly does not randomizing these items affect gameplay in any way?

In all my years of playing Angband, I have never once found either Bladeturner or the One Ring.

I have found Bladeturner-based randarts twice however, because the new base item is much more common than PDSM.

[Pete Mack]

That works of course, but I dont like it. After all, the point of randomizing (for me) is to not know in advance what to expect; with The One and Bladeturner fixed, that makes 2 of the 4 big slots (weapon, bodyarmour, rings) where I know the end. Id much rather have a broken toon ala Aberly now and then than that.

This just is not true. When is the last time you found either Bladeturner or The One Ring? So how exactly does not randomizing these items affect gameplay in any way?

[Estie]

The problem I see with extra attacks on non-weapons is that it distorts the balance between classes. A warrior gets a 50% melee damage bonus in exchange for not having spells compared to a mage. If you have +5 attacks from equipment, this is reduced to ~ +20% while the benefit of spells remains the same. Slays and brands dont have this problem, as their damage increase is proportional to the number of (base) attacks.

I think miyazaki´s suggestion is good; how about this:

1. Leave weapons unchanged.
2. Define new properties, off-weapon_blow and off-weapon_shot. Only non-weapon artifacts can get these. They are independant of any Pval an item might have, like fire resistance.
3. The effect of having the property on a piece is to increase the number of blows by 1 (2 if class = warrior), or number of shots by 1 (2 if class = ranger). They do not stack, so having it on 2 pieces doesnt add another blow; just like fire resistance.