Handaxe Limb Damage Update

edited March 2021 in The Matsuhama Arena

So, given this:

It is Friday, the 5th of March, and the following changes have been implemented:

- Limb damage values on thrown axes has been significantly reduced, per Changes news #40.

I put it to the test, and turns out for non-sentinels (haven't tested Sentinel yet) if you're using the in-line targeting syntax, limb damage on handaxe has not changed at all. It does stand to reason that Sentinels who were using in-line targeting syntax would also see no change here.

While I think this is what was hinted at in Change news #40, the wording of today's post seems to indicate that it has actually been reduced, as in a balance change to the weapon itself, which is not the case - it is just a bugfix for the syntax issue reported a month or two ago. Just hope this clarifies for anyone concerned.

This chart (without any change from before) remains perfectly accurate for all data points gathered through extensive testing (performed today). Can't say it's the exact formula but it's very close.

(Tested in Runewarden with Trans Weaponry, on multiple artefact level handaxes).

Comments

  • edited April 2021

    @Nicola @Caelan

    After some pretty extensive testing on Throw (~20 different weapons including 7 variants of handaxe) on 5 classes with/without proficiency, with/without weaponry/skirmishing/weaponmastery, I gotta say.

    Shit's whack.


    Results are often completely unintuitive, to the point where the whole thing just feels like a giant bug that has been encultured into being a feature. I'm beginning to understand why it can't be shared with us after seeing some of these charts that often make literally no sense.

    Could write a book on this but a few examples below (noting these are NOT actual formulas, they're just basic approximations, but they are quite accurate!) Almost all of the damage values here are accurate to +/- 10 damage, on average across the entire spectrum, noting that the real formulas don't appear to actually be continuous linear functions).


    red: L0/forged handaxe (94 dam)

    green: L2 handaxe (101 dam)

    blue: L3 handaxe (106 dam)

    Noting that forged → L2 is 7 damage increase, while L2 → L3 is a 5 damage increase


    This shows that artefact level NOR damage stat intuitively track with limb damage. L0 to L2 is a tiny jump despite being a big credit cost and also a 7 stat increase, while L2 to L3 is almost triple the increase despite being only one artie level up and only 5 damage increase. Makes no sense.



    Meanwhile:

    Black: Sentinel with L3 axe

    Yellow: Monk (it's the same as Sentinel???)

    Blue: L3 handaxe on Paladin

    Purple: L3 handaxe on Runewarden, runed


    This shows that neither class or skills seem to make very much sense. Monk is the same as Sent. RW is higher than Sentinel. Paladin did more damage than Sentinel (without laguz) over 3k HP opponent HP which eventually becomes even higher than Runewarden WITH rune at very high end. Literally nothing here makes sense.

    Also found more than a few cases where class slot order, cross-class skill levels (skirmishing, weaponry, weaponmastery) were applied when they shouldn't be, and that dragonform limb damage with axe still gets bonuses from lesserform class.




    Here's Runewarden limb damage PER TIME for L3 handaxe (purple), L3 Longsword Slice with buckler/smash (red), L3 Broadsword Rend with buckler/smash (green).

    Noting this isn't a fair comparison because with handaxe you don't have to raze, which in reality gives it another 20-25% boost in value on top of this, in real fights. Also noting that handaxe affliction rate is notably higher than Rend, making Rend literally useless.

    (this is actually true on basically every class - as there is nothing I've found in the game that even comes close to handaxe limb damage in terms of damage/time).









    Handaxe is doing triple the limb damage per time compared to WARHAMMERS while ignoring rebounding and giving venom affs every 1.7s., while also doing more HP damage and more bleeding.


    Can we just, please, make this make sense?

    Why can't we just do something sensical and straightforward, like:


    Throw formula:

    (10% max HP + 4 * weapon damage stat) * modifiers

    where modifiers would simply be:

    proficiency (1/1 at full prof)

    class bonus (1.15 for sents, 1.0 for everyone else)

    Highest of Skirmishing, Weaponry, Weaponmastery skill (get rid of double/triple-skill bonuses). (11/11 at Trans)


    That's it. Greatly simplifies everything without really changing anything. Easy to code, easy to tune, easy to use.

  • Limited testing, but results seemed strange to me as well. Putting it here in case admin want to see it, or someone else has some input? I won't pretend I know any formulas either, but it seems all over the board - especially given that Sentinel is supposed to have modifiers? Full disclosure, my Sentinel damage windows are all really close to like.. 35hp difference, my monk is not so tight, and I don't have any dragon testing yet.

    This 'test' is essentially anectodal and put here just to get some ideas from admin, and other players alike, on what may be effecting/causing things like the runeblade appearing to influence the dragon break if your lesserform class is Runewarden. (Similar to how my monk and sentinel break points are thus far, identical, and my class slot 1 is Sentinel - despite sentinel supposedly being boosted by skirmishing)


    Plain old dragon. RW Lesserform, no runes, no runeblades

    • 6457hp is 7 axes [same as Sentinel]
    • 8012hp is 8 axes (only took 7 axes in monk, unstanced) [7 axes in Sentinel, 7 in monk]

    Runewarden:

    • 8874hp = 8 axes in runie. no runes [same as Sentinel, monk]
    • 8874hp = 8 axes with lagua/l/z runes [same as Sentinel, monk]
    • 8874hp = 7 axes with runebladed axe (eihwaz) [8 axes in Sentinel, 8 in monk]


    Dragon again, but with runebladed axe (RW lesser):

    • 8195hp took 7 axes (note that previously 8012hp took more) [same as Sentinel - not tested in monk yet.]


    7450hp took 6 axes in lesserform, but 7 in dragon. [takes 7 hits in Sentinel, monk]

    8068hp took 7 axes in lesserform AND in dragon. [same as Sentinel, same as monk]


  • You should try using that candycane sword- I'm convinced there's some sort of rng or other modifier at play, because 4k hp was breaking after 8k hp. I've also had tests where one limb breaks before another, back in the day, with Shikudo- left arm less hp than right.


    I'm glad at least it isn't just me finding weird stuff! Thanks for all the tinkering and number crunching.

  • It's entirely possible there's an issue with limb damage somewhere, wouldn't be the first time. I just doublechecked and things seem to be working as intended across classes, but if something in particular stands out you're welcome to report that. I'm not commenting on limb damage mechanics in general, or improvements/how things should be, because those mechanics are over my head and those decisions are above my paygrade. A few notes on how things stand currently, though:

    These linear charts without data aren't all that helpful, because they're approximations without visible data, and having seen at least one of the scatter plots for this sort of limb break data, there are some wild outliers that raise questions about data collection (outliers that aren't at all visible in the linear data). They're also not great because, in practice, that's not how limb damage works. The chart in your first post is a much better comparison, really. I get that they're helpful for data purposes. I'm not at all trying to knock you here, by all means go ahead and handle your data however you like, but it makes it hard to tell how many assumptions are being made in your approximations, and that makes it hard for us to take those complaints and test for issues. We do take complaints about limb damage seriously and test what we can (see: the recent sentinel targetting discrepancy), but the sheer range of circumstances people test under and the secondhand nature of these reports without most of the context make it difficult to actually verify whether there's an issue. Very few of the complaints we receive about limb damage discrepancies are actual issues, most come down to user error or misunderstood mechanics. Most are buried in very long and winding explanations with loads of irrelevant information or misconceptions, which makes it hard to get to the actual issue at hand. If you find a discrepancy, it's way more helpful if you can narrow down that discrepancy to a single example rather than a generalization. This isn't to call anyone out in particular, just a tip to help us actually find out whether there's an issue and handle it accordingly. Most of the bugs that are deleted several times as not a bug, only to be eventually fixed way down the line (something I'm sure many combatants are familiar with), suffer from this sort of problem, where there's a legitimate bug buried under a novel's worth of irrelevant data, misconceptions, and assumptions about how things work, and are dismissed accordingly.

    Tl;dr: Instead of "I think runeblades screw with limb damage, here's an approximation of testing and breakpoints across four damage levels and classes against loads of different people", find a case where you can say "I broke X's leg in 6 hits with an axe, empowered it as a runeblade, and broke the same leg on the same person in the same (non-runewarden) class at 5 hits". There's times where more information is helpful, or where more info can't be provided, but at the very least try to find a concrete example you can point to.

    Sentinels do not now have (nor, as far as I'm aware, have they ever had, certainly not in the last decade or so) a class-specific damage or limb damage buff on axes. I'm not sure where this idea originated, or if it was a thing very very early on, a product of some bad testing, or a quote about the skill that was poorly worded or misunderstood (see the next point about weaponry), but it's a rumour that just won't seem to die.

    A quick read over HELP WEAPONRY will probably explain most other cross-class discrepancies (for throwing axes, at least).

    For @Kog, Candycane sword is going to be screwy because it's a ridiculous outlier that was never intended for limb testing purposes and falls far outside the scope of pretty much every other damage calculation in the game. It wasn't meant to be used for limb testing, nor is it a reliable way to do so, but if you can find some use for that data, be our guest.

  • Alas. I had such high hopes for using it similar to the totem test, but that does make sense- thank you for the answer @Phaestus .

  • edited April 2021

    @Phaestus Thanks. That is sort of the point of my post. To give actual numbers and value points and the gist of how it was checked (so if there is an issue, it is easier for you all to replicate).

    I seem to remember when I first came back (late 2017? ) EVERYONE was saying there was a modifier for Skirmishing - so I took it as gospel.


    Another purpose of my post was to point out tjings like the runebladed axe working in Dragon. Stats dont appear to be a factor. If they are, in one example it seems to support it (breaking quicker than before) and in others it seems to disprove (ie taking MORE hits in dragon than in lesser).


    But I just spent 2 weeks trying to get all my values and the same axe performs differently in dragon and across cladses despite every stat and race/trait staying the same - except monk. I was just wonderibg which was correct really.


    edit: If you eant I can send values, logs, etc. whatever I have.


  • Cross-class discrepancies are expected and intended, as noted above.

    On testing as a dragon (runewarden lesser), on the same target, both before and after making an axe a runeblade offer the same limb damage. Like I said, there might be an issue somewhere, but unless you have a concrete example there's not really much we can do to nail it down.

  • @Phaestus it has been clearly stated that Sentinel has a class-specific "hidden proficiency bonus" by admins. I think @Makarios but I'd have to dig for a while to check.

    Also, Sentinel is supposed to also have a non-hidden bonus that's described in Help Weaponry, where they get 110% skill multiplier with Trans Skirmishing and Weaponry.


    As for the way I graphed it, I actually never graph it this way. I graph the actual data and my functions fit all the data. The approximation lines used here are not useful for calculating limb damage exactly but what they do do, very well, is show the comparison between different class/weapon configurations.

    It's also very, very hard to report a bug when the whole system when working correctly feels like a bug. As I explained above it's quite common for things to simply make zero sense, so it's unclear when we should report a bug.

    For example, why is Paladin not doing the same as monk? If you have Trans Weaponry, handaxe should be identical. There is clearly a hidden mechanic changing it for Paladin. Is that a bug or a feature, it's impossible to tell.

  • Re: Paladin vs monk, classes with weaponmastery gain the same sort of 10% bonus for weaponmastery as sentinel does with woodlore. That could probably be explained a bit better in the helpfile (I thought it was already added, file has been updated accordingly), but it's certainly expected behaviour.

    Re: Sentinel, there's no class-specific buff to either player damage or limb damage for axe throwing.

  • edited April 2021

    That.. yeah.


    so..... "specific issues" are:

    - Runeblades effected dragon.

    - One dragon broke more than another (same target), the only difference was a lesserform.


    - Sentinel is said to have a multiplier but doesnt. If it does, other classes get the same multiplier. @Shecks is right, and a "quick read over HELP WEAPONRY" confirms it (See image below just pulled from a decent resource at https://www.achaea.com/game-help/?id=507 )


    As for specific cases, I presented those too. I am REALLY trying not to sound adversarial, as it is essentially my default state nowadays and text makes it worse. But that ish felt very much like a brush off.


    edit: In the interest of not coming off like a total dick (not that I'm worried if anyone thinks I am), I will caveat my mild-snark with: unless you are saying the confusion comes from HELP WEAPONRY referencing 'weapon effectiveness' and that the 'damage' in there does not include limb damage - which doesn't really make a lot of sense.


  • edited April 2021

    edit: I missed the editing above so I've included some 'quick reads' here.


    From HELP BODYPARTDAMAGE:

    In order to deal out body part damage, which is possible with weapons and with
    some abilities, you must first obtain the TARGETTING ability in the Weaponry
    skill.
    


    This seems to indicate that WEAPONRY would be the go-to resource for finding out how body part damage effects others. Looping back to HELP WEAPONRY. There is no "See also... " so HELP WEAPONRY is it.


    From HELP WEAPONS:

    Now, these are not the only factors that affect how effective you are with a
    weapon. The other factors are covered in the HELP WEAPONRY scroll.
    


    So everything anyone looking to find out about limb damage, gets routed to HELP WEAPONRY, so the 'damage' there should theoretically include limb damage. Especially since it makes mention of potential bonuses to them via class abilities, naming specifically, Skirmishing.


    And that brings us to HELP WEAPONRY:

    "Training" give better effectiveness, including Skirmishing getting said 10% bump (one reason I was a little confused by the scaling DOWN of damage for the in-line targetting fix, because, "We only get a 10% bump already.. ").

    "Proficiency" - maxed

    "Attack Used" - THROW AXE AT <person> LEFT LEG for instance.. Nowhere in HELP DRAGONS or AB DRAGONCRAFT does it say they should be conferred any bonus to throwing weapons on par with any bonuses from other classes.


    (See also: HELP WEAPONS, HELP THROWING)
    
    Weaponry is a skill available to all. Anyone may use any weapon, and any attack
    with that weapon. Whether it works or not is another question! The results you
    get with weapons are strongly influenced by:
    (1) training, which gives you greater "weapons effectiveness," *
    (2) proficiency in that weapon, and
    (3) the attack used (say, using crush with a longsword).
    
    
    (*) In the case of certain class skills training may also increase effectiveness. Specifics are as follows:
     - For those with Groves or Pranks, the higher of the two skills (e.g. Pranks vs Weaponry) will be used, with no benefit from the lower.
     - For those with Skirmishing, the higher skill will be used, with a bonus 10% of the effectiveness of the lower skill.
     - For those with Weaponmastery, learning weaponry will convey no benefit, as the benefits of Transcendent Weaponry are conveyed regardless (only for the purpose of "weapons effectiveness"). Further learning in Weaponmastery will convey up to 10% additional efficacy.
    
    
    Weapons Effectiveness
    ---------------------
    This is a sort of general measure of how well you will use any weapon, and is related, as noted above, to the Weaponry skill (and, for some, the
    Weaponmastery, Pranks, Skirmishing, or Groves skill).
    
    What is affected by your weapon's effectiveness?
    - the chance that any weapon's attack (swing/jab/whatever) is 'a hit.'
    - the chance that your 'hit' is not avoided by your target.
    - the damage done.
    
    


    HELP TARGETTING just says you will do reduced damage with the assumption being more limb damage strategy, less physical removal of sippable health points.


    So there may very well just be some 'missing' item in there that I'm not reading into to see why my Sentinel and Monk get the same bonuses to limb damage from the same weapon, while using the same class, stats, and traits. If so, please do clarify - I mean that sincerely.


  • @Phaestus it has been clearly stated that Sentinel has a class-specific "hidden proficiency bonus" by admins. I think @Makarios but I'd have to dig for a while to check.

    This is different to my experience, where using limb formula derived from data acquired with sentinel works well for other classes. Unless this contradicts the quote from Makarios, sentinel gets a significant boost in dps, limb dps, affs per second from the increased speed from throwing. I dont think this is compounded with a different limb damage / hit.

  • edited April 2021

    @Caelan Sorry for the mix-up, the rest of weaponry is about as terrible as throwing. On closer look, those 10% bonuses shouldn't affect damage or limb damage on throwing, no, so there shouldn't be any cross-class discrepancy (except for runewarden, because of laguz). Testing across several classes (including both Sentinel and monk) shows consistent damage on the classes I tested, so I'm not sure where any other discrepancies you're seeing may be coming from. I checked again just to be super extra sure, and I can assure you that runeblades aren't changing damage in dragonform, so I'm really not sure why you're seeing a discrepancy at all.

    That is to say, I've done just about everything I can do on my end to reproduce either issue mentioned above, and they're not showing. If you can find two people with different health values, where you can break the higher health target in fewer hits with the same axe, file a bug with who you tested on and a log, and we can take a look at that exact case and see what's going on. If you find someone who you can break a limb in fewer hits with a runebladed axe in dragonform than with a non-runebladed axe, file a bug with the axes you used and the person you tested on, and we can look into that exact case. That's what I mean by specific examples. Because I've tested several different scenarios for both different health values and runeblades, and everything looks like it's working as intended.

    Edit: Lyrin is also correct above.

  • I'll email all the actual data.

    There is one fundamental assumption behind all of my testing (because I've verified it as well as I can) which is that limb safe is target agnostic and only cares about their max hp value. If there is ANY factor based on opponent other than max HP then my zillion hours of testing and data has to be thrown in the trash.


    As soon as I'm at a PC I'll share a link to the actual test data spreadsheet with you and the graphs with the actual data points on them.

  • edited April 2021

    @Phaestus - Thanks for that. At least I know I shouldn't have to test on every single class, with and without runes, with and without runeblades. I know we weren't supposed to need to anyway, but...here we are/were. I tested it and got those weird results.


    Could we maybe get clarity in HELP WEAPONRY and/or HELP BODYPARTDAMAGE that 'damage' is not effected by Skirmishing, and neither is limb damage (seems weird to say it is part of the effectiveness, and have effectiveness pretty much defined partially as damage/limb damage). Maybe what IS effected by that 'weapon effectiveness' bonus (or get rid of it if there really has never been a bonus). "If you have skirmishing, you need either weaponry or the proficiency, if you have both a 10% boost to physical damage and accuracy will occur. Note: This does not effect limb damage". Something very explicit so as to avoid the confusion going forward?


    I will msg you the log/link of the log of me testing it. Probably a bit edited since it was over the course of an hour or two, but I will include changing class/dragon and the axing in entirety. (I was manualling it because targetting brings up consequences I didn't want to bring to bear. So ignore the incorrect targetting of 'mhun' and 'homonculus').


  • edited April 2021

    Full disclosure... in doublechecking the log for one point, I realized I was mistaken about another. The person I tested with threw on a bracelet right before one specific test began and I missed it, increasing their hp by about 700.


    Fully concede that I may have, and likely was, wrong on that one point. And I apologize for wasting your time with that:

    "8012hp is 8 axes but with runebladed axe 8195hp took 7 axes "


    Will move discussion to messages and send you the logs so it can be duplicated if an issue exists.

    See @Phaestus this is why no one messes with the legacy code lol. It makes people start asking questions and getting confused with vaguery! I will victim blame you for this one so I don't look bad! 😂


Sign In or Register to comment.