Any point to training a weapon skill once you hit min delay?

[AzuredreamsXT]

Topic, thanks in advance

[Hopeless]

Topic, thanks in advance

As far as I know (and don't quote me in case I am mistaken :p) You get more damage from higher skill with heavy weapons.

[reaver]

Topic, thanks in advance

You get more damage from higher skill with any weapons at all.

Fixed that for you.

Min delay isn't as important as some people hype it to be. It's just a break point - Until you get to Min delay every weapon skill level boosts your damage output by X, then after Mindelay each weapon skill increases it by X/5 or whatever (numbers completely made up). Generally once you hit Min delay you want to train something else for a while until you have enough XP it's worth using on that lower X/5 number, but that's not a hard rule. If you're primarily melee like a Berserker it's definitely well worth training past min delay by the time you reach the Depths.

[Sandman25]

Min delay is extremely important. Even for weapon with the highest base damage.

Fighting 3.

  Code:

Dungeon Crawl Stone Soup version 0.14-a0-2820-gd1528c4
Attack: Troll Arcane Marksman vs. yak (4000 rounds) (2014/03/22/22:23:01)
Troll Arcane Marksman: XL 1   Str 18   Int 9   Dex 9
Wielding: +0,+0 giant spiked club of freezing, Skill: Maces & Flails
yak: HD 7   AC 4   EV 7

Maces & Fl | AvHitDam | MaxDam | Accuracy | AvDam | AvTime | AvSpeed | AvEffDam
         0 |     14.2 |     42 |      54% |   7.8 |   180  |  0.56 |      4.3
         1 |     14.0 |     42 |      54% |   7.6 |   175  |  0.57 |      4.4
         2 |     14.0 |     43 |      56% |   7.9 |   170  |  0.59 |      4.7
         3 |     14.2 |     42 |      57% |   8.2 |   165  |  0.61 |      5.0
         4 |     14.6 |     48 |      60% |   8.8 |   160  |  0.62 |      5.5
         5 |     14.8 |     48 |      59% |   8.8 |   155  |  0.64 |      5.7
         6 |     15.1 |     47 |      61% |   9.3 |   150  |  0.67 |      6.2
         7 |     15.5 |     52 |      61% |   9.6 |   145  |  0.69 |      6.6
         8 |     16.0 |     56 |      63% |  10.2 |   140  |  0.71 |      7.3
         9 |     16.3 |     50 |      64% |  10.5 |   135  |  0.74 |      7.8
        10 |     16.8 |     52 |      63% |  10.6 |   130  |  0.77 |      8.2
        11 |     16.9 |     53 |      64% |  10.9 |   125  |  0.80 |      8.7
        12 |     17.0 |     55 |      65% |  11.1 |   120  |  0.83 |      9.3
        13 |     17.0 |     58 |      66% |  11.3 |   115  |  0.87 |      9.8
        14 |     18.3 |     58 |      68% |  12.5 |   110  |  0.91 |     11.3
        15 |     18.1 |     58 |      66% |  12.1 |   105  |  0.95 |     11.5
        16 |     18.6 |     62 |      68% |  12.7 |   100  |  1.00 |     12.7
        17 |     18.8 |     67 |      67% |  12.8 |    95  |  1.05 |     13.4
        18 |     18.8 |     63 |      67% |  12.7 |    90  |  1.11 |     14.1
        19 |     19.3 |     69 |      68% |  13.3 |    85  |  1.18 |     15.6
        20 |     19.5 |     68 |      69% |  13.5 |    80  |  1.25 |     16.9
        21 |     20.1 |     71 |      70% |  14.1 |    75  |  1.34 |     18.8
        22 |     20.2 |     71 |      70% |  14.3 |    70  |  1.43 |     20.4
        23 |     20.3 |     73 |      69% |  14.0 |    70  |  1.43 |     20.1
        24 |     21.1 |     78 |      71% |  15.1 |    70  |  1.43 |     21.6
        25 |     21.5 |     70 |      73% |  15.7 |    70  |  1.43 |     22.5
        26 |     21.8 |     77 |      72% |  15.7 |    70  |  1.43 |     22.4
        27 |     21.7 |     74 |      71% |  15.6 |    70  |  1.43 |     22.2
-----------------------------------


fighting 20.

  Code:

Dungeon Crawl Stone Soup version 0.14-a0-2820-gd1528c4
Attack: Troll Arcane Marksman vs. yak (4000 rounds) (2014/03/22/22:19:24)
Troll Arcane Marksman: XL 1   Str 18   Int 9   Dex 9
Wielding: +0,+0 giant spiked club, Skill: Maces & Flails
yak: HD 7   AC 4   EV 7

Maces & Fl | AvHitDam | MaxDam | Accuracy | AvDam | AvTime | AvSpeed | AvEffDam
         0 |     14.0 |     43 |      68% |   9.6 |   180  |  0.56 |      5.3
         1 |     13.8 |     44 |      67% |   9.3 |   175  |  0.57 |      5.3
         2 |     14.2 |     43 |      68% |   9.7 |   170  |  0.59 |      5.7
         3 |     14.7 |     44 |      68% |  10.1 |   165  |  0.61 |      6.1
         4 |     14.9 |     46 |      70% |  10.4 |   160  |  0.62 |      6.5
         5 |     15.1 |     50 |      71% |  10.8 |   155  |  0.65 |      7.0
         6 |     15.9 |     49 |      70% |  11.1 |   150  |  0.67 |      7.4
         7 |     15.6 |     53 |      70% |  11.1 |   145  |  0.69 |      7.6
         8 |     16.0 |     53 |      72% |  11.5 |   140  |  0.71 |      8.2
         9 |     16.8 |     53 |      72% |  12.2 |   135  |  0.74 |      9.0
        10 |     16.7 |     53 |      73% |  12.2 |   130  |  0.77 |      9.4
        11 |     17.2 |     57 |      74% |  12.8 |   125  |  0.80 |     10.2
        12 |     17.3 |     60 |      72% |  12.6 |   120  |  0.83 |     10.5
        13 |     17.5 |     59 |      74% |  13.0 |   115  |  0.87 |     11.3
        14 |     18.2 |     62 |      74% |  13.5 |   110  |  0.91 |     12.3
        15 |     18.4 |     64 |      75% |  13.9 |   105  |  0.95 |     13.2
        16 |     18.9 |     61 |      75% |  14.3 |   100  |  1.00 |     14.3
        17 |     19.6 |     68 |      75% |  14.8 |    95  |  1.05 |     15.6
        18 |     19.5 |     69 |      74% |  14.6 |    90  |  1.11 |     16.2
        19 |     19.8 |     68 |      75% |  15.0 |    85  |  1.18 |     17.6
        20 |     19.7 |     71 |      77% |  15.2 |    80  |  1.25 |     19.0
        21 |     20.4 |     74 |      77% |  15.7 |    75  |  1.33 |     21.0
        22 |     21.0 |     79 |      76% |  16.1 |    70  |  1.43 |     23.0
        23 |     21.3 |     80 |      78% |  16.7 |    70  |  1.43 |     23.9
        24 |     21.2 |     83 |      77% |  16.3 |    70  |  1.43 |     23.3
        25 |     21.6 |     81 |      78% |  17.0 |    70  |  1.43 |     24.3
        26 |     22.7 |     80 |      77% |  17.6 |    70  |  1.43 |     25.1
        27 |     22.1 |     80 |      77% |  17.2 |    70  |  1.43 |     24.5
-----------------------------------

[duvessa]

https://crawl.develz.org/tavern/viewtopic.php?f=8&t=10724&p=148247#p148247

The accuracy bonus is 1d(skill)-1 to your accuracy roll - so if your accuracy is 1d20 without skill, with 10 skill it is 1d(20 + 1d10-1). (Your accuracy will never actually be 1d20, that is just an example.)

Until you get to Min delay every weapon skill level boosts your damage output by X, then after Mindelay each weapon skill increases it by X/5 or whatever (numbers completely made up).

It is exceedingly rare in practice for that denominator to be anywhere near as small as 5.

[Abominae]

In cases like Quickblades and Whips, I will eventually train a bit more than minimum delay (anywhere from 14-18 skill). I would never train any of the weapons that require 20+ skill past minimum delay.

Outside of niche cases like xxBe or incredible aptitudes, I don't think it is really worth train past minimum delay until a player is well into extended. I can pretty much always think of a better use of exp (unless of course it is a really low skill weapon)

[Hopeless]

Topic, thanks in advance

You get more damage from higher skill with any weapons at all.

Fixed that for you.

Min delay isn't as important as some people hype it to be. It's just a break point - Until you get to Min delay every weapon skill level boosts your damage output by X, then after Mindelay each weapon skill increases it by X/5 or whatever (numbers completely made up). Generally once you hit Min delay you want to train something else for a while until you have enough XP it's worth using on that lower X/5 number, but that's not a hard rule. If you're primarily melee like a Berserker it's definitely well worth training past min delay by the time you reach the Depths.

Yeah thanks for the sarcastic snark. Needed that to start the day properly. But the reason I specified Heavy weapons is that I've several opinions by slightly better players that the gains from skill in say short blades after min-delay are minimal where as you get a ton of extra damage from heavy weapons skill after min-delay. I recognize this may be wrong but I am not merely pulling it out of the air but from these very forums.

[duvessa]

I don't think correcting incorrect information is "sarcastic snark."

I haven't seen anyone say you get a ton of extra damage from training skill past minimum delay. Ever. Well, until now. You do get more extra damage the higher your base damage and str/dex are, but it will never be very much (as you can see, the difference between 2 or 3 levels doesn't even appear in 4000-round fsim).

[reaver]

Yeah thanks for the sarcastic snark. Needed that to start the day properly. But the reason I specified Heavy weapons is that I've several opinions by slightly better players that the gains from skill in say short blades after min-delay are minimal where as you get a ton of extra damage from heavy weapons skill after min-delay. I recognize this may be wrong but I am not merely pulling it out of the air but from these very forums.

Sorry! I misinterpreted that a bit and thought it was important to emphasize heavy weapons don't have some special case where they alone benefit from post - min delay skill levels. I could have communicated that differently, although it may not have been so visible.

[mikee]

The main reason not to is opportunity cost.

It might help to think of skill training as being in a shop where you buy stats for your character. Your character's weapon is at minimum delay, which means it probably has a bunch of weapon skill already, so another point will cost you about 5000 gold. The shopkeeper offers you +0.5 accuracy and 2% more melee damage (this is a simplification) for 5000 gold. On the other hand, for 500 gold, you can buy a point of AC or EV. For 1000 gold, you can buy 0.5 accuracy, 1% more melee damage, and a few hp. Etc., etc. What would you buy?

[Hopeless]

I don't think correcting incorrect information is "sarcastic snark."

I haven't seen anyone say you get a ton of extra damage from training skill past minimum delay. Ever. Well, until now. You do get more extra damage the higher your base damage and str/dex are, but it will never be very much (as you can see, the difference between 2 or 3 levels doesn't even appear in 4000-round fsim).

The correction (if correct) wasn't snark nor sarcasm. The way in which it done was. I realize this isn't an entirely English speaking crowd so the subtly may have been lost in translation. Basically when someone quotes you on a forum and changes something that's snark and saying "there fixed that for ya" is snark. Not that it was offensive per se, just was an interesting thing to read while I was still waking up.

[Sandman25]

Right, saying correct words is no excuse for being rude.

[Psiweapon]

Oh hey opportunity costs (-:

[Lasty]

The way weapons are structured, increasing skill level is often a bad idea for both low-training and high-training weapons.

Low-training weapons tend to only be good if you're using additive damage (additive brands, slaying bonuses), and training the weapon skill only increases that added damage until you hit min delay. The weapon base damage is so low that training further is fairly useless, even though it's cheap.

High-training weapons have better base damage and often are used with multiplicative brands, making training additional levels give significantly better results, but since you already had to train 20+ levels just to get to min delay, those extra levels are incredibly expensive, and thus are also not really worth training.

There are a few exceptions: demon weapons, lajatangs, dire flails, and a few other weapons have decent base damage and require little training, and so getting extra levels on them might be a reasonable cost/reward ratio sometimes. Also, if you're going for a stabbing-based build, over-training Short Blades just for stabbing bonuses can be a good idea sometimes.

[rchandra]

One consideration to training past mindelay is draining insurance - if you have Drain, you might still be at min delay or not slowed as much. However since those later levels are still expensive, I don't think this will make a difference unless other factors also apply (Trog, cheap weapon, maybe Mf, etc.).

[Rast]

Apts are worth considering too. If your weapon is +3 and Armour and Dodging are both -3, mikee's numbers get a lot different.

[and into]

Apts are worth considering too. If your weapon is +3 and Armour and Dodging are both -3, mikee's numbers get a lot different.

To an extent, yes. But only to an extent, and even then perhaps not in the way you would imagine.

But a total swing of 6 in apts between two things you want to train is extremely rare, and Centaurs are the only thing in the game that have a -3 to both armor and dodging, and yet even centaurs want to eventually train those defensive skills (usually armor for the most part). Basically almost everyone wants to train some combination of armor and dodging regardless of their apts, eventually, and so bad aptitudes for those skills basically exist to put a brake on species that would be overpowered (or even *more* overpowered) if they didn't have bad aptitudes there. In other words the reason for those bad apts is not to discourage people from investing in the skills ever, but to balance things out.

So when it says in the manual that +4 means you learn "twice as fast," this can be rather misleading; it does *not* mean that your skill level, for equal experience investment, will be double that of someone with 0 aptitude, because the cost to get to the next skill level raises really fast.

Quick test on human in wizmode using a manual: Same amount of experience that takes a dude from 0 to 8.1 at normal (0) aptitude will take you from 0 to 11.3 with +4 aptitude (given by a manual). So in this case +4 aptitude gives you 3.2 more skill levels. A Troll training stealth (-4 aptitude) to 6.3 would have his skill level increased just to 9.2 if using a manual, which gives +4 and thus equivalent of 0 aptitude, a difference of 2.9. And a total difference of 4 is a very big swing. (Note: For these tests I used &A to set all skill levels to 0 each time before comparing.)

So yes aptitudes matter, but like most things in Crawl they do not shut off nor completely determine your options or course of development. They simply shift the values around in terms of when it could be reasonable to invest experience in a particular skill, versus what it is best to focus on first. This is a good thing, as it keeps each game of Crawl fluid and open-ended in a way that few games are.

Everything mikee said about opportunity costs remains true even in the cases of extreme aptitudes. So yes it affects it, but (perhaps surprisingly or counter-intuitively) it generally doesn't get *a lot* different.

[Sandman25]

Surprisingly I found one weapon for each training weapon skill past min delay makes some sense. Min delay is at M&F 20 but damage still is increased by about 15% at M&F 27.

  Code:

Dungeon Crawl Stone Soup version 0.14-a0-2820-gd1528c4
Attack: Troll Arcane Marksman vs. yak (4000 rounds) (2014/03/24/21:45:19)
Troll Arcane Marksman: XL 1   Str 18   Int 9   Dex 9
Wielding: +10,+7 giant club of speed, Skill: Maces & Flails
yak: HD 7   AC 4   EV 7

Maces & Fl | AvHitDam | MaxDam | Accuracy | AvDam | AvTime | AvSpeed | AvEffDam
         0 |     15.5 |     40 |      76% |  11.8 |   113  |  0.88 |     10.5
         1 |     15.2 |     43 |      76% |  11.6 |   110  |  0.91 |     10.5
         2 |     15.4 |     40 |      77% |  12.0 |   106  |  0.94 |     11.4
         3 |     15.5 |     42 |      76% |  11.9 |   103  |  0.97 |     11.5
         4 |     15.6 |     47 |      77% |  12.0 |   100  |  1.00 |     12.0
         5 |     16.0 |     45 |      77% |  12.4 |    97  |  1.03 |     12.8
         6 |     16.0 |     48 |      78% |  12.5 |    93  |  1.07 |     13.5
         7 |     16.3 |     47 |      78% |  12.8 |    90  |  1.11 |     14.3
         8 |     17.1 |     50 |      78% |  13.4 |    87  |  1.15 |     15.4
         9 |     16.8 |     48 |      78% |  13.2 |    83  |  1.20 |     15.9
        10 |     17.5 |     51 |      77% |  13.6 |    80  |  1.25 |     17.0
        11 |     17.8 |     51 |      79% |  14.2 |    76  |  1.31 |     18.7
        12 |     18.0 |     51 |      79% |  14.3 |    73  |  1.36 |     19.6
        13 |     17.9 |     55 |      81% |  14.6 |    70  |  1.43 |     20.9
        14 |     18.2 |     61 |      79% |  14.6 |    67  |  1.50 |     21.8
        15 |     18.9 |     57 |      79% |  15.1 |    63  |  1.58 |     23.9
        16 |     19.1 |     57 |      81% |  15.5 |    60  |  1.67 |     25.9
        17 |     19.1 |     63 |      80% |  15.3 |    57  |  1.77 |     26.9
        18 |     19.3 |     65 |      80% |  15.5 |    53  |  1.88 |     29.2
        19 |     19.7 |     61 |      80% |  15.9 |    50  |  1.99 |     31.8
        20 |     19.9 |     64 |      81% |  16.3 |    47  |  2.15 |     34.6
        21 |     20.5 |     61 |      82% |  16.9 |    47  |  2.14 |     36.0
        22 |     20.2 |     65 |      81% |  16.6 |    47  |  2.14 |     35.2
        23 |     21.2 |     70 |      81% |  17.3 |    47  |  2.15 |     36.9
        24 |     21.1 |     69 |      81% |  17.2 |    47  |  2.14 |     36.5
        25 |     21.6 |     67 |      82% |  17.8 |    47  |  2.14 |     37.8
        26 |     22.0 |     73 |      81% |  17.9 |    47  |  2.14 |     38.0
        27 |     22.1 |     78 |      83% |  18.4 |    47  |  2.15 |     39.1


Edit. No need to use wizmode for aptitudes/skills comparison, http://ianberrigan.com/ is much easier to use.

[and into]

Edit. No need to use wizmode for aptitudes/skills comparison, http://ianberrigan.com/ is much easier to use.

I do appreciate the link, I'll keep it in mind, but

THIS CALCULATOR IS VERY MUCH IN BETA AND MAY BE VERY INCORRECT. TAKE ALL RESULTS WITH A LARGE GRAIN OF SALT.

at top of the page makes me wary.

Actually I did put in some numbers to see and the first thing I tried yielded very, very clearly wrong answer. So I would recommend against using that calculator until it is improved.

[Sandman25]

Which numbers did you try? It works for me "Same amount of experience that takes a dude from 0 to 8.1 at normal (0) aptitude will take you from 0 to 11.3 with +4 aptitude (given by a manual)"

[and into]

Which numbers did you try? It works for me "Same amount of experience that takes a dude from 0 to 8.1 at normal (0) aptitude will take you from 0 to 11.3 with +4 aptitude (given by a manual)"

Yes that's correct to within 0.1, but I tried some other numbers with it, and I think it starts getting wonky when you begin dealing with stuff like 2 versus -4 for aptitudes and non-zero starting skill.

For example,

Aptitude of first skill: 2
Current level of first skill: 20
Proposed level of first skill: 26.9
Aptitude of second skill: -4
Current level of second skill: 0

Yields answer: 12.5

According to my wizmode test: 13.5

All other skills set to zero.

[Sandman25]

The page supports integer numbers only so it used 26 instead of 26.9.

[and into]

The page supports integer numbers only so it used 26 instead of 26.9.

Ah okay I see that warning now on the page, alright. Thanks!