minimum delay change

[some12fat2move]

I suggest that minimum delay should be reached at skill 27 for all weapons. This eliminates the breakpoints, or magic numbers, in the skill system that strongly encourage the player to know the formula and do calculations, which is a very bad thing (I think we all agree that magic numbers like this are unpleasantly fiddly and bad to have). Assuming that the formula to determine a weapon's minimum delay stays unchanged, this is what delay would look like. The three weapons there are bardiche, halberd, and spear, the black lines are current delay, the red lines are proposed delay, and the dotted line just marks the spot of skill 27.

There is a very clear benefit to be gained from this proposal because, as I mentioned, I think we all agree that the magic numbers in the current system are problematic. But there is also a potential loss: in the current system, there is an interesting distinction between fast, low-investment weapons versus strong, high-investment weapons. But I believe that we can keep this distinction even with the proposed change. Here's why: A dagger becomes twice as fast at minimum delay, whereas a triple sword becomes about three times as fast. Under the proposed system, neither weapon will reach minimum delay until skill 27, but the triple sword will benefit more from each level of skill. So players would be encouraged to invest more skill into heavy weapons than light ones, and we will see players using heavy weapons having a lot of weapon skill and players with lighter weapons having less weapon skill - like things are now, except without breakpoints. Under the proposed system, there would always be the strategic decision of "Is my weapon worth investing skill in?", you would have to think about it differently depending on your particular weapon, you would of course have to consider your other skill opportunities, and most importantly, the strategic question would ALWAYS be there rather than ceasing to intrigue you once you hit skill 14 or whatever.

Some rebalancing would probably be required in order to implement this proposal (buffing some weapons case-by-case to compensate for the loss in speed would probably be the best way).

[reaver]

I don't think this proposal is fleshed out enough to act on.

I'm not trying to be mean, but there just isn't enough content here. It's two paragraphs saying a common development idea - min delay should be removed - and a single suggestion on how to possibly implement it, which is so common a branch has already implemented it, although it was never merged. If programming is constructing a building and design is making blueprints for that building, this is like pointing to a plain lot where everybody want to build a bank and suggesting somebody build a bank there with two stories.

Some rebalancing would probably be required in order to implement this proposal (buffing some weapons case-by-case to compensate for the loss in speed would probably be the best way).

This is what a proposal would need. A system balanced when it's proposed, with numbers to support it. I don't think any proposal has a chance to get picked up unless every weapon is given its new numbers, fsims and spreadsheets are made proving that the best weapon is still debatable and that different weapons are good for different characters. I always thought this series of articles would be useful for somebody trying to rebalance the weapons. (Note: Start at part one).

[Mulzaro]

Nerf all weapon speed to shut down weapon skill magic numbers. They all reach min delay at skill 27.

Note that thorough discussion (thread) showed that a lot of careful weapon stat re-balancing would be needed with this to not make two-hand weapons grossly overpower one-handed weapons.


From Crawl development wiki

[Klown]

Would suck to need 27 just to use a demon whip.

The 'magic numbers' and what skill you need to hit min delay should be more clear in-game, though. Maybe even popping up your 'manage skills' screen saying 'you have minimum delay with your current weapon, would you like to switch skill training now?'

[crate]

Having magic numbers for weapon skill is bad but the one thing it does well is it makes 1h weapons improve more quickly that 2h weapons. There was a proposal here last year or so which had some dev support that basically made it so going from skill 0->27 in any weapon gave you 1 extra attack per turn with that weapon. That succeeds in one way: it makes weapon skill's effect smooth, which is a good thing. The problem is that it makes 2h weapons improve more quickly, at any skill level, than 1h weapons, given the damage/delay values for the weapons in the game.

If you want to do something similar I think you are going to also have to revamp weapon stats from scratch, because right now crawl weapons break down like this: 1h weapons are bad but get better quickly at low skill (for instance qblade is 5 dam/7 delay at skill 0, but maxes out at skill 8). 2h weapons are good but get better slowly (lajatang is 16/14 at skill 0, great mace and great sword are 17/17 and 16/16 respectively). I have thought a bit and did not come up with anything that would both eliminate magic numbers and retain the dynamic that currently exists. Possibly you are smarter than I am! If that's the case then great, let everyone know how to do it. The proposal in the OP does not do this; it makes big 2h weapons get better more quickly than small ones (I don't want to do the math to compare e.g. qblade to 2h stuff), which is a problem since the big weapons are already better at skill 0 (see the slopes of the lines in the graph). Alternatively you can redo all the weapon stats, but hopefully you can see how that would be difficult to get into crawl.

[duvessa]

A lot of people neglect that you would also be making most weapon base types redundant.

[some12fat2move]

Okay so you guys are saying basically: What seems like a sensible solution is to make heavy weapons be bad, compared to lighter ones, at zero skill, but scale better with training to become better. This makes sense and introduces decisions. Making weapon delay go linearly down all the way to 27 makes heavy weapons scale better, but that alone is not enough because they would still be better than lighter ones even at zero skill. So all that's needed is to make light weapon base types better.

Well that doesn't sound so hard. Doing so case by case might result in, for instance (you needn't bother reading through this list I'm just giving you an example):

  Code:

Weapon / Dam / Acc / Delay

Dagger / 5 / +10/ 100%
Quick Blade / 6 / +8 / 60%
Short Sword / 6 / +11 / 100%
Sabre / 7 / +10 / 110%

Falchion / 9 / +9 / 130%
Long Sword / 11 / +8 / 140%
Scimitar / 12 /  +5 / 140%
Demon Blade / 13 / +6 / 120%
Great Sword / 16 / +0 / 160%
Triple Sword / 19 / -3 / 190%

Spear / 7 / +9 / 110%
Trident / 10 / +6 / 130%
Demon Trident / 12 / +6 / 120%
Halberd / 13 / +4 / 150%
Glaive / 15 / -1 / 170%
Bardiche / 18 / -6 / 200%

Hand Axe / 8 / +9 / 130%
War Axe / 11 / +7 / 150%
Broad Axe / 13 / +3 / 160%
Battleaxe / 15 / -1 / 170%
Executioner's Axe / 18 / -6 / 200%

Whip / 6 / +9 / 110%
Club / 5 / +4 / 130%
Hammer / 7 / +8 / 130%
Mace / 9 / +7 / 140%
Flail / 11 / +6 / 140%
Morningstar / 13 / +4 / 150%
Eveningstar / 15 / +3 / 150%
Demon Whip / 11 / +9 / 110%
Dire Flail / 14 / +2 / 130%
Great Mace / 17 / -3 / 170%
Giant Club / 20 / -4 / 170%
Giant Spiked Club / 22 / -6 / 180%

Staff / 5 / +10 / 120%
Quarterstaff / 11 / +7 / 130%
Lajatang / 16 / +0 / 140%

However as crate mentioned, that is a lot of work to look over and probably isn't spot on for all the weapons and would need a lot of adjustment and checking and that would be a pain in the ass and because of that it is extremely likely that no one would ever bother to implement it. So it would be better to take a formulaic approach, like, for instance:

all weapons with a delay of less than 16 auts (meaning 160% if I understand auts correctly) get +1 to base damage
all weapons get base accuracy increased by (20 - auts)

That looks pretty fair to me, it would make light weapons tons more accurate than heavy ones and I think probably better at low skill?

[duvessa]

What seems like a sensible solution is to make heavy weapons be bad, compared to lighter ones, at zero skill, but scale better with training to become better.

Which is nearly impossible to do with your proposed system. For a great sword to be worse than a falchion at 4 skill but better at 27 skill, it would need to have a base delay of about 400%. Or have like -18 base accuracy, in which case any one of having a to-hit enchantment or having slaying or wearing body armour would throw it way off balance.

Unless you're proposing that every weapon have a different scaling factor in which case you're replacing an easy to understand magic number with one that's harder to understand and near impossible to even see in the game, which hardly seems like an improvement, let alone an improvement that's worth having to rebalance something that affects almost every character in the game.

[MDvedh]

Actually, I like the current system.
Min delay means something like "You need to be _that_ skilled to use your weapon most effectively."
The bad part is you need to use spoilers to know that.

How about an entry in the weapon info screen. Something like "You need 12 skill levels in Maces&Flails to use this weapon effectively."

[HenryFlower]

75% of all of this would be solved by a magic power like UI for weapon speed and damage:

E.g., demon whip vs great mace at 5 skill (making the numbers up)

  Code:

Weapon       Speed              Damage Per Hit  Average Damage
Demon Whip   ######******|      #####**|        foo
Great Mace   #*****|            #######***|     bar

Other 25% would be solved by better damage bonus for weapon skills (which could be easily indicated in the damage/hit graph.

[galehar]

There was a proposal here last year or so which had some dev support that basically made it so going from skill 0->27 in any weapon gave you 1 extra attack per turn with that weapon.

The dev support probably was because it was a dev proposal
Here it is. In addition to moving the min delay to skill 27, the proposal also makes weapon speed linear with skill instead of weapon delay.

Okay so you guys are saying basically: What seems like a sensible solution is to make heavy weapons be bad, compared to lighter ones, at zero skill, but scale better with training to become better.

Exactly. That's what I went for with my 2nd revision, but I didn't fleshed it out much. I didn't even put it on the wiki. A shame since it was promising IMO.

[galehar]

What seems like a sensible solution is to make heavy weapons be bad, compared to lighter ones, at zero skill, but scale better with training to become better.

Which is nearly impossible to do with your proposed system. For a great sword to be worse than a falchion at 4 skill but better at 27 skill, it would need to have a base delay of about 400%. Or have like -18 base accuracy, in which case any one of having a to-hit enchantment or having slaying or wearing body armour would throw it way off balance.

Maybe making big weapons worse than small ones at skill 0 isn't really possible. But it doesn't matter. The point is that we can make them worse than they are now (by increasing the base delay) if they catch up and end up at the same power level as now at high skill level.

edit: oh, I forgot to tell about the split of the spell analogy. Here it is if you want to pursue this fascinating discussion.

[Leafsnail]

Actually, I like the current system.
Min delay means something like "You need to be _that_ skilled to use your weapon most effectively."
The bad part is you need to use spoilers to know that.

How about an entry in the weapon info screen. Something like "You need 12 skill levels in Maces&Flails to use this weapon effectively."

I really agree with this. The only part of the philosophy that the status quo really goes against is the "no spoilers" rule, and that can be easily fixed (by adding a description to the weapon and possibly a message when you reach minimum delay). "Do I want to spend more XP in order to use a better weapon" seems like a perfectly meaningful choice to me.

I can see that complicated breakpoints involving multiple factors are bad but if they can be communicated and understood easily I don't see the problem.

e: I also like the idea of giving the player some information about how accurate/ damaging their weapons are for them.