Gloves vs Claws

[megawidget]

Would you wear randart +3 gloves {Str+5 Int+2} as an unarmed ghoul?
On a related note, how much value would you give slaying bonuses over AC?

[Sandman25]

I think the best approach would be to use fsim.
For normal gloves I usually don't use them early game but use them later because damage loss is huge early game (up to 35% as far as I remember) and small (5%) at UC 27. Yet your gloves have +Str and it can play important role too. "AC vs slaying" depends on AC/EV/monster/play style etc. so there is no absolute answer IMHO.

[megawidget]

I fsim'd to establish that I lose about 1.5 damage as a result. I'm guessing that +4 net AC is worth it, so I'm wearing them, but I don't actually know what I'm doing. :p

[Sandman25]

I fsim'd to establish that I lose about 1.5 damage as a result. I'm guessing that +4 net AC is worth it, so I'm wearing them, but I don't actually know what I'm doing. :p

You can use &f to see both offense and defense vs a melee monster. Run it twice with/without gloves and you will see what is better.

Edit. Or just post results here and we will help you to decide what is better.

[WingedEspeon]

unless it is super early I would wear them. The int may also be slightly useful as ghouls int is so low stat death is a minor concern.

[tasonir]

Generally speaking I'd use them, although posting the results of fsim would be useful, I'm surprised that the +5 str doesn't make up for the loss of damage. But basically you're taking them for the 4ac. As for damage, it probably depends on how high your UC is, as sandman mentioned, claws 1 isn't a huge amount of damage at higher skill levels. Str is approx 3% damage, so 5 str is 15%, which is less than the 35% early game figure and more than the 5% maxed UC figure.

[Hurkyl]

I fsim'd to establish that I lose about 1.5 damage as a result. I'm guessing that +4 net AC is worth it, so I'm wearing them, but I don't actually know what I'm doing. :p

Keep in mind that while losing 1.5 damage when you average 30 is not a big deal, losing 1.5 damage when you average 3 can be catastrophic!

[megawidget]

For anyone who still cares about this thread, fsim confirms this -- +5 Str gauntlets result in the same damage as ghoul claws at UC27, so there's no reason not to wear them. My UC is at 23, so I'm only losing maybe one point of damage, which is easily compensated for just by AC.

[Lasty]

Definitely wear the gloves. Even at low levels, the loss of damage isn't that important when you consider how much 4 AC and better Str for armour help you. It's not like barehanded UC doesn't work. And if you get them early enough, you could instead pick up a decent weapon anyway.

I disagree with the advice to fsim this (and every other choice). It's available for people who want to compulsively run numbers, but it's not mandatory. Often it'll lead you to misleading conclusions, since people who run fsim rarely run it across the spectrum of possibilities required to get a full picture.

Side note: how +5 str impacts your damage depends on not just weapon skill, but also Fighting skill.

[Sandman25]

I don't think running fsim is ever a bad advice, it will not kill your character.
My last UC character used GDA with 19 Str and Armour 9, do you have any idea how much damage I lost comparing to robe? I suggest fsiming it, I suspect you will be surprised.
Another thing that surprised me a lot was Fighting, I didn't expect difference between Fighting 15 and 20 is so big in GDA.

[Lasty]

I don't think running fsim is ever a bad advice, it will not kill your character.

It can be a good way to confirm a mistaken conclusion through inadequate testing, which can make it suboptimal within a given game. What's far worse is that it can give people the impression that the game is easier or more fun when you run fsim constantly, and that's not true at all -- at least for many.

You like to play in a way that relies heavily on fsim, but that's just one play style. Convincing newbies to constantly run fsim will cause some of them to quickly stop enjoying the game.

[Sandman25]

It can be a good way to confirm a mistaken conclusion through inadequate testing, which can make it suboptimal within a given game. What's far worse is that it can give people the impression that the game is easier or more fun when you run fsim constantly, and that's not true at all -- at least for many.

You like to play in a way that relies heavily on fsim, but that's just one play style. Convincing newbies to constantly run fsim will cause some of them to quickly stop enjoying the game.

Can you please provide example of "confirm a mistaken conclusion through inadequate testing", I don't understand what you mean here. Weapon of poison/antimagic?
I am not convincing players to run fsim constantly, I am convincing them to learn how to fsim (and use wizard mode in general) just to get correct answers in 5 minutes instead of getting wrong answers after creating threads "which weapon should I use", "which armour should I wear" or "which skill should I train".

[pickled_heretic]

fsim is a great tool and I used to use it more, but I do want to suggest that it doesn't do a good job of accounting for spikes (or droughts), or getting hit by more than 1 monster, or getting chewed by jackals while also getting smote by an orc priest, or getting smacked around by a blink frog while you make for a staircase, so at the end of the day you almost always need to make a heuristic assessment and the only way to get better at making heuristic assessments is to make more heuristic assessments, which involves not using fsim.

[Sandman25]

fsim shows max damage also.

[megawidget]

To clarify, at 27 UC + Fighting, gloves of strength do the same damage, roughly, to a titan as ghoul claws. This has been tested with a level 27 GhGl who has only been pumping dex.
At 24.9 UC + 21.1 Fighting I lose 1-1.5 points of damage.

My randart +5 Str gloves are definitely worth it.

[Berder]

fsim is a great tool and I used to use it more, but I do want to suggest that it doesn't do a good job of accounting for spikes (or droughts), or getting hit by more than 1 monster, or getting chewed by jackals while also getting smote by an orc priest, or getting smacked around by a blink frog while you make for a staircase, so at the end of the day you almost always need to make a heuristic assessment and the only way to get better at making heuristic assessments is to make more heuristic assessments, which involves not using fsim.

Well, fsim really helps sometimes, as data you consider to make your heuristic assessment. It's not the only factor to consider, but it's a useful one. I would say this gloves/claws decision is one such case where fsim can help out. You do have to be smart and interpret the results well, with respect to how much damage you're dealing vs. how good your defenses are vs. how fast your attack is.

(Psst - remember that time when melee damage was doubled for everyone and nobody noticed for two weeks during a tournament?)

[tasonir]

If you do a bad test and get bad results, the solution would be to do a better test, not to completely abandon testing at all. Sure, accounting for all the factors is difficult, but the more you test changing each factor, the more you'll understand about combat mechanics.

Honestly I'm surprised the strength only equals claws at 27 UC - I would have expected them to break even around 15 UC or so, when the +2 from claws would only be ~11% damage. But I guess that +2 being multiplied by fighting makes it a bigger deal? But your 18 base damage is also multiplied by fighting, so I'm not sure why that would change anything. Is the strength bonus only applied to base damage before fighting?

[pickled_heretic]

the ultimate test that accounts for all the factors is actually playing the game.

[tasonir]

It is nearly impossible to tell how much damage you're dealing by playing the game, as demonstrated by thousands of players playing for days with double damage, and most of them not noticing, some of them having vague notions of being strong, and only a handful actually figuring out what happened.

I have dozens of wins before the bug and I four-streaked during the tournament and didn't realize I was doing 100% more damage. This ghoul going from 40 to 41.5 damage CANNOT be determined in game.

[pickled_heretic]

and that's my point - if you don't notice the damage, the ac is probably better.

you would have definitely noticed double damage if it had toggled during an actual game.

[megawidget]

If you do a bad test and get bad results, the solution would be to do a better test, not to completely abandon testing at all. Sure, accounting for all the factors is difficult, but the more you test changing each factor, the more you'll understand about combat mechanics.

Honestly I'm surprised the strength only equals claws at 27 UC - I would have expected them to break even around 15 UC or so, when the +2 from claws would only be ~11% damage. But I guess that +2 being multiplied by fighting makes it a bigger deal? But your 18 base damage is also multiplied by fighting, so I'm not sure why that would change anything. Is the strength bonus only applied to base damage before fighting?

This is because the claws bonus is not slaying, but is part of base damage. Since stat-based damage scaling is linear, increasing strength suffers from diminishing returns. Therefore, an already high-strength character like a ghoul may benefit more from keeping claws.

  Code:

        potential_damage = using_weapon() || wpn_skill == SK_THROWING               
            ? weapon_damage() : calc_base_unarmed_damage();                         
                                                                                     
        potential_damage = player_stat_modify_damage(potential_damage);             
                                                                                     
        damage = random2(potential_damage+1);                                       
                                                                                     
        damage = player_apply_weapon_skill(damage);                                 
        damage = player_apply_fighting_skill(damage, false);                         
        damage = player_apply_misc_modifiers(damage);                               
        damage = player_apply_slaying_bonuses(damage, false);                       
        damage = player_stab(damage);


[tasonir]

and that's my point - if you don't notice the damage, the ac is probably better.

you would have definitely noticed double damage if it had toggled during an actual game.

But my point equally applies to AC - You won't notice the difference that the 4 ac makes to your actual damage received either. But, AC is printed on your screen, so you can definitely tell "my ac number is higher now" and thus you'll be more likely to pick it. If it showed "AC: 23, damage: 42" vs "AC:27, damage: 40" then you'd be able to make an informed decision about which is better. But it only shows the AC. You could argue that 2 damage off attacks is a lot and should be noticeable, but that only applies to attacks which aren't blocked/dodged, or already reduced to 0 by your AC. So it's 2 damage off about 1/4 or 1/3 of attacks, depending on circumstances. It's not something you're going to notice unless you're collecting statistics.

I'm not sure if the bug toggling mid-game or at the start of a character really makes that much difference. I certainly noticed when it was turned off, but that's because I was aware of the bug and knew it was being turned off, so I looked at my damage closely.

[Berder]

and that's my point - if you don't notice the damage, the ac is probably better.

you would have definitely noticed double damage if it had toggled during an actual game.

It toggled in trunk, so people who had saved games from before double damage would have updated to the double damage version. They didn't notice. That also includes anybody who had lag and refreshed the browser, updating to the new version.

[pickled_heretic]

I suppose all these people were specificially observing their damage output because they were expecting a change? (like, say, a change to their equipment)

[tasonir]

Now you're moving the goalposts. Just because you expect a change when you change equipment doesn't mean you'll have powers of detection to detect a 30 times smaller difference (100% damage vs about 3%).

[pickled_heretic]

how am I moving the goalpost? if I am expecting a change, I am more likely to observe one, even if one doesn't actually occur. See, I can link random bullshit on a wiki too! And yes, I know this doesn't really support my support my original point any more than yours, so maybe I should run a fsim and see which point makes more sense?

[Sandman25]

Without gloves:
You hit Ettin!
You hit Ettin but do no damage.
You hit Ettin!!!
You hit Ettin.
You hit Ettin!!
You hit Ettin!
You hit Ettin!!
You hit Ettin!
You hit Ettin.
You hit Ettin!!!

With gloves:
You hit Ettin!!
You hit Ettin!
You hit Ettin!!!
You hit Ettin but do no damage.
You hit Ettin!
You hit Ettin!
You hit Ettin!!
You hit Ettin!!!
You hit Ettin!!

PS. I am playing with mod which displays damage as number and even then it is extremely hard to see even +30% average damage.

[pickled_heretic]

Without gloves:
You hit Ettin!
You hit Ettin but do no damage.
You hit Ettin!!!
You hit Ettin.
You hit Ettin!!
You hit Ettin!
You hit Ettin!!
You hit Ettin!
You hit Ettin.
You hit Ettin!!!

With gloves.
You hit Ettin!!
You hit Ettin!
You hit Ettin!!!
You hit Ettin but do no damage.
You hit Ettin!
You hit Ettin!
You hit Ettin!!
You hit Ettin!!!
You hit Ettin!!

Exactly.

[Sandman25]

Exactly.

What do you mean? I hope you don't think it helps to compare weapons.

[dowan]

Without gloves:
You hit Ettin!
You hit Ettin but do no damage.
You hit Ettin!!!
You hit Ettin.
You hit Ettin!!
You hit Ettin!
You hit Ettin!!
You hit Ettin!
You hit Ettin.
You hit Ettin!!!

With gloves:
You hit Ettin!!
You hit Ettin!
You hit Ettin!!!
You hit Ettin but do no damage.
You hit Ettin!
You hit Ettin!
You hit Ettin!!
You hit Ettin!!!
You hit Ettin!!

PS. I am playing with mod which displays damage as number and even then it is extremely hard to see even +30% average damage.

So obviously gloves give you more '!', 2 more, to be exact. However, no gloves allows for the unique possibility of '.'.

After checking these ! and . against the chart in the manual where one sees the value of '!' and '.', I see... hmm... no mention at all.

Spoiler: show

Also, to throw in my 2 cents in the fsim vs don't fsim debate: In the worlds of medicine, science, and generally important things like that, when one wants to know the answer to a question, it's generally agreed that the best possible method to get that answer is to do some kind of trial. Generally as many trials as possible. If the trials return numerical results, generally the average of said results is the best way to get an accurate answer. If the trials were set up properly to actually answer the intended question, this is actually considered the best possible way to answer it, rather than say, asking four or five random people on the internet.

[Pollen_Golem]

I didn't know there was a unique called Ettin.

[Berder]

Well, moving the goalposts is sometimes just a misunderstanding. One person says something, a second person gives their claim an uncharitable interpretation that lets them refute it, then the first person explains what they really meant and the second person claims the first is moving the goalposts. In this case the fallacy is really that the second person misunderstood the first.

Sometimes moving the goalposts is really about conceding the point. One person says something, a second person refutes it, and the first person concedes that point and then makes a different claim instead. There's no fallacy here. The first person admits his first claim was wrong, but maybe his second claim is correct (or not). If the second claim does turn out to be correct, that has no bearing on the veracity of the first claim. The fallacy would only be if the first person went back to asserting his first claim despite evidence.

I would say pickled_heretic has a point that people weren't specifically watching for the damage change. However, the fact that they didn't notice is still very interesting and speaks to how hard it is to see changes in damage.

[Sandman25]

So obviously gloves give you more '!', 2 more, to be exact. However, no gloves allows for the unique possibility of '.'

It is useless info. You can get more "!" in feedback for a worse weapon because it is random.
fsim uses 4000 attacks and still shows some anomaly sometimes.

[tasonir]

and that's my point - if you don't notice the damage, the ac is probably better.

you would have definitely noticed double damage if it had toggled during an actual game.

It toggled in trunk, so people who had saved games from before double damage would have updated to the double damage version. They didn't notice. That also includes anybody who had lag and refreshed the browser, updating to the new version.

You claimed it would be noticed if it happened to an in-progress game. Berder stated that it first occurred in trunk in active games, and when they reloaded they had suddenly double damage. You then claimed this didn't matter and that you'd be able to tell the difference after equipping the gloves because you'd be trying to find a difference. This part here is moving the goalposts.

While it's true you're going to be trying to find the difference after changing equipment, the fact that this difference is 30 times smaller than one that went unnoticed, means you won't be able to notice the difference. Other posters tried to point this out by pasting actual damage messages that you'd have to try to interpret. Actual data/statistics are the only way you're going to notice a change this small. I don't believe anyone tracks the number of hits for 0 damage, hits for "." damage, hits for "!" damage, etc, from BEFORE they find an item, and then change gear, and continue tracking it afterwards. Or they find an item, and now spend 20 minutes going back through their log and tallying how many !'s they had.

While it's always possible I misunderstood something, imho "you would have definitely noticed double damage if it had toggled during an actual game." is a fairly simple, direct claim.

[Sandman25]

I don't believe anyone tracks the number of hits for 0 damage, hits for "." damage, hits for "!" damage, etc, from BEFORE they find an item, and then change gear, and continue tracking it afterwards. Or they find an item, and now spend 20 minutes going back through their log and tallying how many !'s they had.

Even that is not enough. You can sum total damage before and after but there is no guarantee that they will show real picture because they are random.
Before "5, 2, 7, 8", total 22
After "3, 8, 9, 7", total 27
Is after better? We don't know for sure, maybe we were just lucky.

[pickled_heretic]

by "toggling during an actual game" I didn't mean "on the same persistent character, that may or may not have been played months ago" I meant "toggling during an actual game."

[savageorange]

by "toggling during an actual game" I didn't mean "on the same persistent character, that may or may not have been played months ago" I meant "toggling during an actual game."

It toggled in trunk, so people who had saved games from before double damage would have updated to the double damage version. They didn't notice. That also includes anybody who had lag and refreshed the browser, updating to the new version.

Most webtiles/SSH servers update multiple times a day. It's not precisely clear how regular 'very regularly updated' is, but I would guess hourly (by contrast with the entry for crawl.berotato.org, which says 'updates daily at 0800 GMT'). So yes, toggling during an actual game would have occurred.

[Arrhythmia]

by "toggling during an actual game" I didn't mean "on the same persistent character, that may or may not have been played months ago" I meant "toggling during an actual game."

WHAT ARE YOU TRYING TO SAY

[and into]

Well, the word "toggle" does imply a change that one has actively induced, like flipping a light-switch, rather than a change that has occurred without user input. So, when you equip something, you are aware that you are changing something about your character, and you thus might pay closer attention to how your damage input/output seems to have changed. (Even when you are looking for it, however, it can still be difficult to see changes that are nevertheless potentially significant. Particularly for new players.)

[savageorange]

Well, the word "toggle" does imply a change that one has actively induced, like flipping a light-switch, rather than a change that has occurred without user input.

It was actively induced, but not by players. Playing webtiles/SSH trunk is pretty much a tacit acceptance that this is likely to happen during your game, though usually the magnitude of the effect is much smaller.

Anyway not to be sidetracked by small disagreements, I agree with the rest of your post. I would personally comment that even when I'm paying attention, I don't necessarily notice 'damage per turn doubled/halved'. It's more like 'damage dropped' or 'damage raised'. A pretty coarse sense, really, I think Crawl encourages it to be coarse by not providing numbers.