Forum Settings
       
Reply To Thread

Puppet Damage numbers, fSTR, etc (MASSIVE WoT)Follow

#1 Jul 20 2012 at 3:10 PM Rating: Good
is Happy on Friday!
*****
12,443 posts
ok, using my findings from this thread as my basis, I'm going to try and discern a reasonable equation for the puppet's damage value. The reason this is PURELY THEORYCRAFT is because this sh*t is @#%^ing obnoxious to test. For those not sure what I mean: Ranged Attacks work differently than melee attacks, especially for the puppet; Sharpshot's ranged attack always does the same damage relative to a target's DEF and VIT. In other words, if it hits something for 400, it will always hit something for 400, there will be no variance, unless either it's ranged skill, str, or dex (see the other thread, pup ranged damage is a little wonky) change, or the target's def or vit change. Melee attacks don't work like that. And to further complicate/simplify sh*t, puppet ranged attacks CANNOT critical. EVER.

Now, first I want to clarify, when I say "damage" in this thread, I'm talking about the invisible damage value for puppet melee attacks, like the one we have for our h2h attacks (h2h damage = ((floor(hand-to-hand skill * 0.11) +3) + weapon damage) (ie: the +18 damage on hades sainti)), and interestingly, the number for our puppet's ranged damage uses a nearly identical equation (auto ranged weapon damage = (floor(automaton ranged skill * 0.11) * 3), with the only difference being the skill source and a single sign. Now, using these equations, and the average damage rating on a number of different weapons, I'm going to attempt to find a reasonably accurate potential equation for automaton melee skill.

Theoretically, the best way to test Valoredge's weapon damage would be the same way people used to use to figure out if they had broken (read: activated) their destroyers (or any other of the latent: Damage+x Crit+6% weapons); Sneak Attack vs level -1~0 mobs outside of one of the starter cities (usually rabbits in east or west ronfaure); the idea being that you get a moderately sized group of critical hits and find the highest value, then attempt to use that to establish the "crit cap" for a given hit. The problem with this is that the puppet can't use sneak attack, so you're pretty much forced to use Dynamo and a constant stream of thunder maneuvers to attempt to force your puppet to crit as often as possible. Now, with 3 thunder maneuvers dynamo gives an astounding +9% crit rate (can you feel the sarcasm?), which combined with the base 5% and the surely capped dDEX gives your puppet a 29% crit rate. You could augment a mirke wardecors for an extra 2%, but that'd be a huge waste of time. Regardless, 29% crit rate severely slows down the process, and I honestly can't even rule out that the dex from three thunder maneuvers would influence the results, given how dex impacts ranged attacks. This issue is why I'm purely theorycrafting at this point, because I figure it will be easier to do the math first, and then try to prove it wrong. For the record, I will also be putting up numbers for the other frames as well, since part of me feels valoredge may get a slight boost outside of its extra skill.

Now before I get to number crunching, I need to cover something. Much to my annoyance, I am not 99 yet (stuck at 95 until I get LB10 done), and I can't @#%^ing find any listings of the relevant automaton stats at 99; so I need one of you to list off the base stat values for all four frames at 99 (I want str/dex/vit/agi/mnd/int/chr, only str and dex should be relevant, but I want all 7 for each frame so I can @#%^ing post them somewhere and not need to hunt them down again)
Here's what I have written down from before, that I'm going to use, but would like confirmed:

Sharpshot:
STR: 92
DEX: 96

Valoredge:
STR: 102
DEX: 92

Harlequin:
STR: 98
DEX: 93

Additionally, if someone could be so kind, I would greatly appreciate the following, entirely for reference: Str/Dex values for level 99 pup/war, mnk/war, pld/war, war/sam, War/nin, and str/dex/agi for 99 rng/war; preferably Hume, but if you're not hume, just make sure to say what race you play for clarity. I'm asking for these just to make sure I have accurate numbers, because alla's character sheets tend to be somewhat off for stats after like, level 60 Smiley: glare but they're not really vital, I just need them for reference to know what weapons to compare sh*t to (and thus, help me determine puppet fSTR and Weapon Ranks, because puppet fSTR works very differently than player fSTR, by which I mean sharpshot's fSTR at 99 isn't automatically capped against level -1~0 bunnies in west ronfaure, it required 2-3 fire maneuvers to see a cap.)

Alrighty then, let's establish our baselines:
A+ Skill at 99: 424 (Monk H2H)
A- Skill at 99: 417 (Valoredge Melee; Sharpshot Ranged)
B+ Skill at 99: 404 (Puppetmaster H2H)
B- Skill at 99: 388 (Harlequin Melee)
C+ Skill at 99: 378 (Sharpshot and Stormwaker Melee)
C Skill at 99: 373 (Soulsoother and Spiritreaver Melee)

Mnk Base Damage: 49
Pup Base Damage: 47
Sharpshot Base Ranged Damage: 135

Potential Increases: (for h2h skill, I'm only including realistic sh*t that would actually get used)
Mnk: H2H skill
+7 (Faith or AF3+2 gloves): Base Damage +1 (50)
+14 (Faith and AF3+2 gloves): Base Damage +2 (51)
+16 (Capped Merits): Base Damage +2 (51)
+23 (Capped Merits and Faith or AF3+2 gloves): Base Damage +3 (52) (note: +30 doesn't increase damage, mnk is left 1 skill short of the next damage tier, unless they use sh*ttier gear which kills the benefit of the +1 damage gain)

Pup: H2H skill:
+5 (AF3+2 hat): Base Damage +0 (unfortunately, the AF3+2 hat is 1 skill shy of pushing us over, for those pups who can't afford h2h merits for some retarded reason)
+7 (Faith): Base Damage +1 (48)
+12 (AF3+2 hat and Faith): Base Damage +1 (48)
+16 (Capped Merits): Base Damage +2 (49)
+21 (Capped Merits and AF3+2 Hat): Base Damage +2 (49) (same deal, though, +5 still can't push us over on its own)
+23 (Capped Merits and Faith): Base Damage +2 (49) (however neither can the +7 from faith)
+28 (Capped Merits, AF3+2 Hat, and Faith): Base Damage +3 (50) (of course, this is not necessarily optimal, since it means 1: no cirque necklace, and 2: no rancor collar, both of which are, realistically, the stronger choice for TP most of the time)

Puppet Ranged Skill:
+10 (Merits or AF3+2 pants): Base Damage +3 (138) (remember, ranged damage goes up by 3, not 1, because of the equation)
+20 (Merits and AF3+2 pants): Base Damage +9 (144)

Puppet Melee Skill: (note: excluding Wayang Kulit mantle for a reason; it's not worth owning. Unless my math says otherwise, at which point I'll retroactively add it back in)
+10 (Merits or AF3+2 pants)
+20 (Merits and AF3+2 pants)

Now, because they're multiplied by 0.11, both h2h skill damage and puppet ranged damage follow a tiering of +1 dmg every 9 skill levels for h2h, or +3 dmg every 9 levels for auto-ranged, however, do note that for auto-ranged, it is not +1 every 3, it is +3 every 9, and that for some reason, unlike h2h, which starts logically at 10 skill (10*.11 = 1.1, floored = 1+3 = 4), ranged skill does not, and instead has an extra 4 skill buffer added to it, such that it starts at 14 skill. Essentially, it uses a weird variant of the floor function (to refresh your memory, Floor( means that if a decimal exists in the answer, you round down to the nearest whole number; ie, both 5.001 and 5.999 will round to 5 if they're floored. Floor is SE's favorite function, and it's present in nearly every important FFXI equation in some form or another, like how TP uses a modified floor( and rounds to the 10th place (13.723 and 13.799 will round to 13.7). Thus, for auto-range damage, it doesn't round down exactly, since the math would suggest the tiering would start at 10 skill, but in practice, it started at 14, so there's still some work I need to do to figure out what causes that, but the equation I found held true for what I did test (I checked every damage tier from level 50 up), so I'm assuming it has something to do with a funky floor function early on. It works for all the important levels (75+) anyway, so it's good enough for now.

ANYWAY

For a base line, let's compare the damage values for shaprshot with those of for level 99 ranged weapons:
level 99 Sharpshot, raw (no skill increase)
Dmg: 135 Delay: 360 (based on 9.3tp/hit, because if you used the 20 second ranged attack wait, sharpshot would have a delay of ~1881, which is all sorts of unrealistic)
Appearance: Crossbow, thus, Marksmanship
A note about ranged damage and delay: ranged damage is the combination of both weapon and ammo damage ratings, and but ranged delay only combines the two for TP. That said, given that we're using the TP-based delay, I will be assuming ammo is part of the equation, sort of.)

Ok, firstly, Longbows and Guns are right out, because the typical delay of both weapons is entirely too high. That leaves us with Crossbows and Short bows. There's a good bit of variation from crossbow to crossbow, and SE hasn't added a true short bow since the level cap went up, so there aren't any high level references

Typical Crossbow Delay: 280/288 (ignoring heavy crossbows like O.Bow)
- Special Mention: Repeating Crossbow: Dmg: 17 Delay: 144 (This is the crossbow used to make sharpshot)
Typical Short bow Delay: 360 (For true short bows, anyway, this is ignoring longbows that use short bow models, and various special case bows)

Now, a comparison of a crossbow and a short bow of equal levels has the shortbow having a little bit more damage. For example:
Kaman vs Arbalest (this is really the only perfect match I can find at a reasonable level, next best matches are 2-4 levels apart) (for reference, Sharpshot's skill cap at level 50, 153, would give it 48 damage)

Next, let's compare ammo:
For a level 50 bow, you'd probably be using either Scorpion Arrows or Bodkin Arrows, while for a level 50 crossbow, you'd likely be using either Mythril Bolts, Black Bolts, or Darksteel Bolts (if you assume you're a little over level 50). You should be able to see a bit of a trend here, Arrows are typically lower in both delay and damage relative to bolts, which is the opposite of their weapons.

For the sake of the argument, let's assume level 52, using darksteel bolts vs level 52 using bodkin arrows. Level 52 cap would up sharpshot a damage tier as well (technically, it's level 51 that does that, but still). Thusly, you get the following:
Arbalest+DSbolts: DMG: 79 Delay: 288 (shooting)/480 (TP)
Kaman+Bodkins: DMG: 60 Delay: 360 (shooting)/450 (TP)
Sharpshot: DMG: 51 Delay 360 (TP)

When you consider that the base weapon used for sharpshot was the Repeating Crossbow, which has an incredibly low damage rating, you can make an analogue for sharpshot's weapon: a Crossbow that has ~160 delay, with bolts that have ~200 delay, and damage such that the equation's damage rating could realistically be split 1/3 for the crossbow's raiting, and 2/3 for the ammo (so at level 52, it's be a 17 dmg crossbow with 34 dmg bolts). Keep in mind that Sharpshot gets a siginificant boost as far as fSTR is concerned (and that its ranged attacks are effectively a weapon skill with str and dex mods), compared to a player, so the reduced damage rating makes sense.

Now, the melee results will be slightly different, however, I think it's probably safe to assume that a similarly HIGH fSTR cap is in play. It is worth noting, however, that even for players, ranged weapons have a significantly higher fSTR cap (and lower floor) relative to 1h/2h/h2h weapons, so that may also come into play. In my tests, I found that to cap dSTR (note: dSTR is different than fSTR; dSTR is shorthand for the difference between your STR and your target's VIT, and it does have a natural cap for players, through the fSTR cap; ie, a point after which further increasing the difference will cause no change) vs a target with 1 VIT, sharpshot needed 102 str. Before I continue, I'm gonna add the stat values for rng from alla's character thingy (so PLEASE tell me if they're incorrect, because they really don't look right to me)
Hume, 99rng/49war:
STR: 94
DEX: 94
AGI: 94

and Sharpshot:
STR: 92
DEX: 96
AGI: 102 (assumed, I can't remember so I'm basing it off of VE's STR)

For the rng, I'm going to use Moros Crossbow with Dark Adaman Bolts as my reference, so that both are using crossbows; thus, Dmg: 68+67 (135) (delay isn't relevant here), and for sharpshot, I'll assume capped 99 skill, but no extra skill, so 135 dmg (I want the numbers close, more than anything else).

I'm going to see if I can't find a rng with a moros crossbow, and ask them if they'd be so kind as to shoot off some d.a. bolts against bunnies with Choke cast on them (to ensure vit=1) in ronf, with 102 str, to get a perfect comparison of damage, but until then, let's compare what I can:

Moros Crossbow has a weapon rank of 7, which puts its fSTR2 (fSTR is for melee weapons, fSTR2 is for ranged weapons) range at -14~30;
For sharpshot, STR >101 DEX96 vs bunnies with 1 VIT yielded consistent 786 damage hits, so if I could compare that to whatever a rng with 102 str and 135 total weapon damage gets (as well as knowing what weapon, so that I can account for weapon rank and determine their fSTR cap), I could quickly figure out the difference in fSTR between players and sharpshot.

However, for the moment, I do know that sharpshot's fSTR is, at the very least, somewhat influenced by it's level (fSTR capped off at 84 str at level 75), and is seemingly independent of auto-ranged skill. Essentially, what I noted was as follows:
- fSTR capped the same at level 96 with 416 skill skill (thus, 135 damage) as it did at level 99 with 417 skill (still 135 damage)
- fSTR at level 99 was consistent from 417 to 437 skill, which covers 3 damage tiers and dmg+9, which, if used pure dmg -> weapon rank -> fSTR, would cause the cap to go up.
- fSTR at 75 with 269 skill capped with dSTR at 84 (so str=85)

Now, testing my theories as best I can: (using 102 str, 96 dex, and 135 damage)
If sharpshot's ranged damage worked the same way as a regular player's:
- If the equated damage rating is all considered the WD:
---Weapon Rank would, theoretically, be 15 (significantly higher than any other ranged weapon in game, including level 99 yoichinoyumi and annihilator; for reference, level 99 Ragnarok has a weapon rank of 15 as well)
--- from a WR of 15, the fSTR2 range would be -30 to 46
--- with an easily assumed pDIF cap of 3.0 (boosting r.atk did nothing, and all damage results were multiples of 3), one can assess that prior to the dex and str mods, with fSTR capped, the damage would be 583, so 203 damage came from STR/DEX mods
--- If we treat the ranged attack as a quasi-ws, meaning we account for WSC, but ignore fTP... working backwards from 786 damage, 135 comes from skill, 46 from fSTR, and 81 from WSC. If I assume 0.85 for alpha... that means 96 is the floored sum of the stat mods... but that's also the raw dex value... and since str seems to cap I guess that would mean it's only a dex-based WSC, of 100%. So this actually fairly cleanly translates to the Damage = ((WD + fSTR+ WSC) * pDIF) formula Smiley: dubious I didn't expect that. I'll have to check some other numbers from my testing...

hmmmm, it works cleanly here, but it seems like the alpha value may be slightly different, since the other numbers seem to be off by 1 dmg (the calculation's expected results are 3 damage lower than what I actually got).. yea, playing with it more, the alpha definitely @#%^s with it, because it only works for my level 99 numbers. I'll come back to this.

- If part of the equated damage rating is considered "ammo"
Seems very unlikely now, after that, but I'm still going to try it, using 1/3 and 2/3 values, so 45 and 90, which give WR of 5 and 10 respectively
5 and 10 would put the fSTR2 at -10~26 and -20~36; if 1/3, that would have the dex mod needing to add 101 damage, while at 2/3, it would need to add 91. 101 damage is unrealistic, because it would put the dex mod over 100%, but 2/3 would put it at 95% if that pesky Alpha value is ignored... hmmm.... but that doesn't work below 99 either, and is even further away from the mark, so let's go back to the first idea.

Now, it is entirely possible that the puppet's "alpha" value is 0.85 at level 99, but different before that. If I take a look at my level 75 numbers...
str 85 dex 73 caps fSTR2, and results in 579 damage. For those tests I had 269 skill, which is 87 base dmg, which would put fSTR2's cap at 34, however this would leave 72 damage for dex to make up. the WS alpha at 75 is 0.83, which doesn't even remotely allow that, even at a 100% mod, so I think at this point I can pretty safely assume that pup ranged attacks do NOT use alpha values as we know them.

I'm gonna go play with these numbers, and leave this post up to hopefully get some response as to the stats I need verified while I do. This thing'll be a loooooooooong term project.

edit: Ok, playing with numbers, I had forgotten something; the equation to determine fSTR itself. That equation, and the one that determines range, given the how much STR it took to hit the fSTR cap, just don't add up.

Basically, it comes down to this; dSTR is 101 (102 - 1); Using the already known player ranged fSTR2 equations, and Sharpshot's damage's supposed weapon rank, I get two very different answers.The one I've already listed above is 2(Weapon Rank +8), which is where I got 46, since weapon rank is (Damage/9). However, the equation for determining your actual fSTR2, not just the cap, is ((dSTR)+4)/2, which gives an fSTR of 52.5. because I know 101 dSTR is what hits the cap, this means that puppet ranged damage simply cannot use player fSTR2 formulas. Thus, I tried melee fSTR, and got this result:
Max = (weapon rank +8); thus 23, while the forumla is ((dSTR)+4)/4, which comes out to 26.25. Closer, but still noticeably different in practice. This means that Sharpshot's ranged attacks use a third set, essentially an fSTR3 (or if you prefer, fSTRp for fSTRpet, but I'm hesitant to use that until I see whether puppet melee isn't something ELSE)

Now, based on my numbers, a 100% dex mod seems impossible, because it would require a higher fSTR value as the puppet's level gets lower. Further, I know the puppet's fSTR is influenced by it's level, but I don't know whether that's purely because of the difference in base damage due to skill, or if it's because the puppet's actual level comes in to play (I do have the means to test this out, though, because I rarely ever use sharpshot, so my ranged skill is a good bit undercapped, and I'm going to log on and play with it after I finish this edit). However, at level 99, weapon D accounts for 135 of 262 damage, leaving 127 damage for 102 STR (101 dSTR) and 96 DEX to account for, and at level 98, losing 1 DEX caused no change in damage, while at level 96, losing 3 dex caused a change in damage of 1 (783 damage instead of 786), meaning that at level 96, STR and DEX account for 126 damage, instead of 127. Now, I'm going to go get some numbers for what I can, and see how they compare.

edit2: Alrighty, my skill atm is 266, which, conveniently, puts me exactly at the level 75 dmg tier (87). At level 95, I have 88 STR and 92 DEX, and I confirmed that the fSTR cap I have now is definitely higher than it was at 75, and it's in line with what I observed at 96 (ie: still within 101 dSTR), however, as I'm looking over my numbers, there's definitely some wonky **** going on that I didn't notice before that if I had, would have automatically ruled out the player fSTR equations, however, I realized that it's a lot ******* harder to test this **** on the live server, because I don't have readily available Cirque Guanti +1 AND +2, and the ability to quickly change my levels.... ergh....

Edited, Jul 20th 2012 10:03pm by Jinte
____________________________
Theytak, Siren Server
LOLGAXE IS MY ETERNAL RIVAL!

Reiterpallasch wrote:
Glitterhands wrote:
Am I the only one who clicked on this thread expecting actual baby photos [of Jinte]? o.O

Except if it were baby photos, it would be like looking at before and afters of Michael Jackson. Only instead of turning into a white guy, he changes into a chick!
#2 Jul 21 2012 at 6:55 AM Rating: Excellent
The Great Pumpkin
*
54 posts
Sorry, I usually try to keep the automaton pages updated on the wiki, stats are now updated on the frame pages:
Harlequin Frame
Sharpshot Frame
Valoredge Frame
Stormwaker Frame

Edit: Had to update the test client. Here's the Job str/dex/agi for a 99 hume
	STR	DEX	AGI 
PUP/WAR  86	 91	 89 
MNK/WAR  92	 91	 80 
PLD/WAR  94	 83	 77 
WAR/SAM  94	 89	 87 
WAR/NIN	 94	 90	 90 
RNG/WAR	 86	 86	 94 
 


Edited, Jul 21st 2012 9:18am by Zijder
#3 Jul 21 2012 at 11:30 AM Rating: Good
is Happy on Friday!
*****
12,443 posts
Zijder wrote:
Sorry, I usually try to keep the automaton pages updated on the wiki, stats are now updated on the frame pages:
Harlequin Frame
Sharpshot Frame
Valoredge Frame
Stormwaker Frame

Edit: Had to update the test client. Here's the Job str/dex/agi for a 99 hume
	STR	DEX	AGI 
PUP/WAR  86	 91	 89 
MNK/WAR  92	 91	 80 
PLD/WAR  94	 83	 77 
WAR/SAM  94	 89	 87 
WAR/NIN	 94	 90	 90 
RNG/WAR	 86	 86	 94 
 


Edited, Jul 21st 2012 9:18am by Zijder


so many Smiley: cookies for you Smiley: inlove
____________________________
Theytak, Siren Server
LOLGAXE IS MY ETERNAL RIVAL!

Reiterpallasch wrote:
Glitterhands wrote:
Am I the only one who clicked on this thread expecting actual baby photos [of Jinte]? o.O

Except if it were baby photos, it would be like looking at before and afters of Michael Jackson. Only instead of turning into a white guy, he changes into a chick!
#4 Jul 22 2012 at 11:29 AM Rating: Good
Sage
***
2,098 posts
good luck.
figuring out pet damage calculation is a massive undertaking and its hard.
____________________________
Server: Bahamut
90BST|90BLU
LS: MONSTERSINC

Choh Moui | Rongo-Nango | Lhu Mhakaracca | Lungo-Nango | Nyumomo
--Beastmaster Forever--
#5 Aug 03 2012 at 9:55 AM Rating: Good
is Happy on Friday!
*****
12,443 posts
Once I can get on the test server to play with mighty strikes, I'll finally be able to get some concrete numbers on this sh*t. Check back in a couple hours, folks, I may even be able to get an idea on fTP/WSC for some of our puppet's WS.

edit: scratch that time frame, apparently I have to completely reinstall the test server, which includes downloading three multi-gig ZIP files. even getting ~500KB/s, this is gonna take me another 3-5 hours...


Edited, Aug 3rd 2012 11:31am by Jinte
____________________________
Theytak, Siren Server
LOLGAXE IS MY ETERNAL RIVAL!

Reiterpallasch wrote:
Glitterhands wrote:
Am I the only one who clicked on this thread expecting actual baby photos [of Jinte]? o.O

Except if it were baby photos, it would be like looking at before and afters of Michael Jackson. Only instead of turning into a white guy, he changes into a chick!
Reply To Thread

Colors Smileys Quote OriginalQuote Checked Help

 

Recent Visitors: 23 All times are in CDT
Anonymous Guests (23)