Issues on current master

[Pete Mack]

Gwarl--as Derakon says, off-weapon brands are highly problematic. The trouble is particularly acute with the very common Glaive of Pain, which does big damage already, and MASSIVE damage with any kind of slay or brand. But the issue isn't limited to that. Things like Eorlingas and Claris become more flexible when they work against most animals, and weapons of Gondolin become much more powerful.

[Derakon]

That being said, it could be worth experimenting with having gear that gives a flat damage bonus per blow against specific kinds of targets. Like "+5 vs. ogres", that kind of thing. The problem with our current set of slays/brands is that they have varying impact depending on what kind of weapon you have, so even a nominally-early-game pair of gloves might be endgame-quality equipment if it has the right slay. But you could have a pair of gloves of +2 Fire Damage that would be useful early while still tapering off properly in the late game.

[Gwarl]

Well that's why I suggested using 'weak' brands for off-weapon brands, i.e. only doubling the base damage rather than tripling it.

[Estie]

That being said, it could be worth experimenting with having gear that gives a flat damage bonus per blow against specific kinds of targets. Like "+5 vs. ogres", that kind of thing. The problem with our current set of slays/brands is that they have varying impact depending on what kind of weapon you have, so even a nominally-early-game pair of gloves might be endgame-quality equipment if it has the right slay. But you could have a pair of gloves of +2 Fire Damage that would be useful early while still tapering off properly in the late game.

Giving Paur* brands would indeed turn them into endgame equipment, but that can be taken account of by making them deeper and rarer. It then becomes a choice of branding your weapon versus getting the flat +8/+10 per blow vs everything plus stats from the other top tier gloves.
Minor bonuses like "+5 vs. ogres" are not going to matter. Your going to use that exactly when you have no other option for the glove slot. Free action, universal damage boost and stats are all better. I dont think trying to wiggle in more small properties before free action is a good idea. I commonly get my first pair of gloves with some kind of property by the first or second towntrip.

There has been the tendency to give items a little bit of everything, which removes choice in the gear assembling process. You just pick "the next upgrade", like in an mmo, instead of making hard choices between, say, str or dex of an amount that matters noticably.

Getting power upgrades in Angband is the best thing, but increasing the amount of times you get upgrades doesnt improve the experience, for the same reason that the lowest scoring type of football is the most popular football.

[PowerWyrm]

There are clearly a few issues still with randarts - rings being based on =Escaping much more often than expected, for example. I need to go away and think carefully about my statistical assumptions...

Ok I generated a game with randarts, suicided and generated randart.log to have a look at randart generation for rings. And I see two issues:

1) The base item chosen depends on the power of the tval. If that power is too high, this happens:

Creating Light
********** Evaluating for base item power ********
Artifact index is 34
the Band 'Aewenna' <+1>
jewelry power is 4, power is 4
2 for searching skill, power is 6
3 for light, power is 9
3 for ignore acid, power is 12
1 for ignore electricity, power is 13
3 for ignore fire, power is 16
1 for ignore cold, power is 17
6 for light resistance, power is 23
FINAL POWER IS 23
Base item power 23
Power too low!

Creating Speed
********** Evaluating for base item power ********
Artifact index is 34
the Band 'Aewenna' <+4>
jewelry power is 4, power is 4
80 for speed, power is 84
3 for ignore acid, power is 87
1 for ignore electricity, power is 88
3 for ignore fire, power is 91
1 for ignore cold, power is 92
FINAL POWER IS 92
Base item power 92

Basically the code tries to pick up a sval with the highest "base power" possible. Looking at the rings, the best choices are rings of speed and then... rings of escaping. No wonder why you get a lot of these...

2) Base power is sometimes miscalculated. Look at this:

Creating Free Action
********** Evaluating for base item power ********
Artifact index is 34
the Ring 'Aewenna'
jewelry power is 4, power is 4
12 for see invisible, power is 16
3 for ignore acid, power is 19
1 for ignore electricity, power is 20
3 for ignore fire, power is 23
1 for ignore cold, power is 24
FINAL POWER IS 24
Base item power 24

See invisible on a ring of free action???

My guess is that the problem with rings of escaping comes from both issues:

Creating Escaping
********** Evaluating for base item power ********
Artifact index is 50
the Ring 'Halith' <+4>
jewelry power is 4, power is 4
80 for speed, power is 84
-9 for impaired mana recovery, power is 75
3 for ignore acid, power is 78
1 for ignore electricity, power is 79
3 for ignore fire, power is 82
1 for ignore cold, power is 83
FINAL POWER IS 83

You get +80 from the speed bonus, and -9 from some impaired mana recovery that isn't on the ring. Here you should get a big penalty due to the AFRAID property (currently -20, should probably gets buffed up) instead.

And yes, FREE_ACT is SEE_INVIS + 1 and AFRAID is IMPAIR_MANA + 1, so there's an off-by-one error somewhere in the power calculation for object properties.

[Estie]

The problem with the ring of escaping is that it defies the rules for artifacts. A ring with "no melee" property gets assigned power -90 (or whatever), but if you add (+10,+10), (+2 attacks) to it, the power is still -90 and decreasing that value doesnt fix the issue. It´s not a matter of flawed statistical assumptions.

I still think taking the ring of escaping out of artifact generation is the best option, or, indeed, taking it out of the game completely.

[Nick]

I still think taking the ring of escaping out of artifact generation is the best option, or, indeed, taking it out of the game completely.

I don't think that will be a problem if all types of rings are allowed, rather than just Speed and Escaping.

And thanks, PowerWyrm, for that excellent analysis - I was starting to think that the power limits might be at least part of the problem, but that makes it really clear. I now have plenty to think about.

[Derakon]

Giving Paur* brands would indeed turn them into endgame equipment, but that can be taken account of by making them deeper and rarer. It then becomes a choice of branding your weapon versus getting the flat +8/+10 per blow vs everything plus stats from the other top tier gloves.

Making them deeper and rarer then means that the only early/mid-game artifact gloves will be Cammithrim.

Minor bonuses like "+5 vs. ogres" are not going to matter. Your going to use that exactly when you have no other option for the glove slot. Free action, universal damage boost and stats are all better.

So tweak the numbers until they are potentially worth using. The point is to have an alternative to the current "puts a multiplicative brand on your weapon" and "gives you flat bonus damage per blow against everything" off-weapon combat boosts.

Hell, we could make Cambeleg be "+8 damage vs. evil" and the Paur* gauntlets be "+12 fire/frost/etc. damage". There's lots of room to play with different approaches.

There has been the tendency to give items a little bit of everything, which removes choice in the gear assembling process. You just pick "the next upgrade", like in an mmo, instead of making hard choices between, say, str or dex of an amount that matters noticably.

I don't disagree with this, but I think it's been mostly a problem with the early-game artifacts (thancs and paurs), which got boosted to make them more "competitive" with midgame arts. Which mostly just meant that if you found an early game artifact early, then you got a free ride to the midgame.

[Huqhox]

Hell, we could make Cambeleg be "+8 damage vs. evil" and the Paur* gauntlets be "+12 fire/frost/etc. damage". There's lots of room to play with different approaches.

I like that idea an awful lot. You could get *Enchant Weapon* to occasionally add a random +1 damage brand as well as the flat damage. The plain multipliers have always seemed very unsubtle and in some ways are opaque to new players (ie.e that different brands/slays have different multipliers and how they fit together)

Would the monster list need rebalancing for this though? Surely flat damage is going to result in utlimately less damage output for the player, certainly in late game. Although that might stop every class essentially turning into buffed warriors at end game and meleeing everything

[Gwarl]

I actually think flat additive bonuses to damage are the wrong approach - which is why gameplay-focused variants like Oangband and Sil remove them completely. I find brands/slays as they are now a more interesting property. I think Vanilla should retain flat additive bonuses, perhaps with some adjustment, simply for tradition's sake, but I don't think we should be introducing any more.

The area of combat which needs the most attention IMO is blows per round calculations. Once the calculations are well understood and perhaps rewritten (I might volunteer to do this if nobody beats me to it) a considered approach to reworking damage calculations could be possible.

I feel like much of angband's traditional combat mechanics could be salvaged by carefully recalculating the BPR table.

[Gwarl]

I like that idea an awful lot. You could get *Enchant Weapon* to occasionally add a random +1 damage brand as well as the flat damage.

Additive bonuses aside, I think applying brands to a weapon using a scroll is a great idea, and I think I saw some unused functions related to this in the code already?

Another idea:
Instead of having off weapon brands at all, make an effect/activation which applies the brand to all the player's attacks for the duration.

[t4nk]

Assertion failed!

Program: ...
File: obj-pile.c
Line: 73

Expression: obj -> prev == prev

Would be great if someone were able to tell how to reproduce that... I looked into angband's pile management (insert, excise, etc) and was unable to find any bugs.
Nick, maybe pile_check_integrity() should dump all information about pile (op and all objects involved) to a file on error? Someone then will encounter this bug, post the file and that should give at least some rough idea about the bug.

[PowerWyrm]

The problem with the ring of escaping is that it defies the rules for artifacts. A ring with "no melee" property gets assigned power -90 (or whatever), but if you add (+10,+10), (+2 attacks) to it, the power is still -90 and decreasing that value doesnt fix the issue. It´s not a matter of flawed statistical assumptions.

If you put -90 on the AFRAID flag, you'll get a base power for rings of escaping of -2, which will reject the choice of rings of escaping for 99.999% of randarts except a few heavily cursed ones, since the choice of tval will depend on that base power. So you'll never get (+10,+10) or +2 attacks on any of them.

[Derakon]

If you put -90 on the AFRAID flag, you'll get a base power for rings of escaping of -2, which will reject the choice of rings of escaping for 99.999% of randarts except a few heavily cursed ones, since the choice of tval will depend on that base power. So you'll never get (+10,+10) or +2 attacks on any of them.

That doesn't change Estie's (accurate) point that it doesn't matter how good the combat bonuses are on a randart if it also has a flag that completely prevents you from participating in combat. The AFRAID flag should just outright prevent addition of to-hit, to-dam, blows, off-weapon brands, etc. to the item.

In principle I wouldn't mind seeing the occasional Ring of Escaping-based randart; they could potentially be great gear for a mage, ranger, or priest if they had good benefits to compensate for the AFRAID flag.

[Ingwe Ingweron]

Using a flavored staff in an attempt to id-by-use. Get the message, "you have no curses to remove", but, the staff remains unidentified. Obviously, a staff of remove curse.