Fatality Belt Latent Testing

[Karbuncle]

Fatality Belt ... whats a Fatality a Finishing Move ... hmmm.. whats that thing DNC get when they use steps ...oh yeah.. Finishing Moves..

idk I would think it would activate if you have 5/5 Finishing Moves.

This is my thoughts!

That was the first thing that came to mind when i read about the item in update patch.

Then again i was thinking it could be active with 1 Finishing move and its "enhanced" by having 2/3/4/5!

Wish i had one to test it

[Malacite]

or maybe it only activates when the target is near death?

http://www.youtube.com/watch?v=r3R0XLErJC8

[NynJa]

or maybe it only activates when the target is near death?

http://www.youtube.com/watch?v=r3R0XLErJC8

Suggested allready

did you try mob being in red hp (sub 25%)?

[Nishida]

I highly doubt SE is thinking Mortal Kombat when making gear.

[Narol]

I highly doubt SE is thinking Mortal Kombat when making gear.

But they do think Highlander when they make synthesis materials.

[Onorgul]

But they do think Highlander when they make synthesis materials.

The US localization team are not the same people as the item developers.

[Demogorgon]

Randomly throwing this in here as well as it might just be kill shot weaponskill.....
Not sure if this is up to the testing standards of bg but, got bored and went sam/thf to east ron last night..... normal sa melees were all 311-317 damage regardless of belt/no belt... not a big enough range that anything has changed... then did 30 gekkos in whatever gear I had with me + no belt/belt/warwolf belt (for comparison)

No Belt: (71+38 str)
Low: 826
High: 868
Avg: 851

Fatality Belt: (71+38 str)
Low: 844
High: 888
Avg: 859.6

Warwolf Belt: (71+43 str)
Low: 856
High: 887
Avg: 865

Can post each individual ws if needed, but it was 30 gekkos per test. Seems to be about a 3% increase in critical damage, so yeah.... outside of multihits (assuming the latent is proper, its also going to be hard to time a ks that lets ever hit land on say drakes/de/penta/etc) this is a pretty garbage belt.

[LinktheDeme]

Might be low % HP WSes and not only kill shot?

[eva00r]

I believe, largely from its name and the gekkos above, it is a kill shot.

The text in JP version is also called "Fatal Belt", with English sound written in Japanese text.

[ringthree]

Might be low % HP WSes and not only kill shot?

If it was kill shot it would be completely worthless.

[Jimie]

If it was kill shot it would be completely worthless.

It'll mean your THF can more easily Sneak Attack killshot while main-handing THF's knife on full moon!

[Psion]

Still think it's moon phase judging by the name of the NM and the meaning of it's name...

[Harbls]

How about when closing a skill chain.

[Sho]

If it was kill shot it would be completely worthless.

Maybe for multihit WSes if only a little. I'd like to believe that the game would not be able to judge when the "kill shot" would land however.

Maybe the latent is while under a negative status effect.

Edit: Lol, nevermind the above.

[Shamaya]

I just wanna look at a few of the samples and kind of check for sufficiency.

Kegsay, the OP, used Beast Slayer (D42+12fStr). With that, given our current pDif information, his max should be 54*~3.15=170 and min 54*~2.65=143

The way I do my tests (or rather, tell other people to do theirs) on level 1 mobs is to continue the sample until it's roughly close to sufficiency. By sufficiency, I mean the point at which the highest # in the sample, divided by the lowest #, is roughly equal to the '1-hander pDif crit cap max/min'. This would be,

pDif max/min (3.15/2.65)
1.189

For Kegsay's tests, the max/min's are,
Without Belt (max HP/MP)
1.167
Though this doesn't matter since via the math we can know how much dmg he should do without the belt anyway.
With Belt (max HP/MP)
1.156
With Belt (Weakend and Red HP; done with Self-Destruct)
1.101
With Belt (Orange HP)
1.152
With Belt (Yellow HP)
1.174

None of these samples are very close to sufficiency (real pDif max/min) of ~1.189. But it's a start. I wouldn't tease OP though, because I really appreciate his effort and want to know the latent myself. What we can conclude though is,
(1)Not one of the samples has a dmg max higher than 170 (which is the predicted max assuming the latent is not triggered). If even one sample did, it would give us reason to believe that the latent was active, without even needing a high 'sufficiency rating' for the sample.
(2)None of the samples are very close to sufficiency max/min of ~1.189. This means it's wholly possible that if he continued the samples, one of them could yeild a number higher than 170. This is least likely for 'with belt, yellow HP' since 170 / 144(the lowest point in his spread) = 1.18, or very close to sufficiency.

This is to be expected since the samples are only 10 each. Even though I have a feeling nothing will turn up from continued testing, if youre reading this OP, I think you should continue the samples until you reach suffiency (max/min's of ~1.189), at least for red HP if anything.

* * *

Continuing on, Demogorgon did a test as well. Here are his suffiency ratings,
No Belt: (71+38 str)
Low: 826
High: 868
max/min: 1.050
Fatality Belt: (71+38 str)
Low: 844
High: 888
max/min: 1.052
Warwolf Belt: (71+43 str)
Low: 856
High: 887
max/min: 1.036
I can see why he'd want to do a warwolf belt sample. We could make use of the data if we do some calculations, but otherwise it can't be compared to the other two samples.

For 2h pDif, the max/min at crit cap is a little different, and appears to be,
3.15/~2.725, or ~1.156.
So, not one of his samples is sufficient on its own. However, if you take his highest sample w/ the belt and divide by the lowest w/ no belt, you should get a suffiency rating (or max/min) of 1.156 or lower if there is no effect. If you get something higher, then it either means that the pDif information is wrong, or the latent is active. If we do this, we get,
888/826 = 1.075. Still insufficient

Likewise, nothing can be concluded from these samples. We don't know if the latent is or isn't active.

* * *

Meanwhile, I got a PM from Wolfheart of Phoenix. Here are his samples,

Did a little fatality testing. Not tons, but little at a time on that. WSing level 1 mobs is not the pinnacle of exciting. And anyway, it's probably enough to for some kind of damage pattern.

after 50 SAWS in 123 dex. Wearing blau. /sam. And not wearing fatality belt: average Evisceration damage was 577 on level 1 mobs. Low of 521. High of 613.

Same gear/sub/weapon Wearing fatality belt: Average evis damage was 569. Low of 511. High of 608.

The difference between these two seems to be nonexistent from that small sample. Same stats, and crit+ belt scores slightly lower. Going to go out on a limb and guess this does not activate on WS, unfortunately ><. I'm still curious what the latent is though. Mostly just curious how retarded SE made the latent.

I checked wiki, and the pdif page is in a shambles. No hard pdif rules have been nailed down, so I can't calculate what my min-max damage should be.

And what I wrote him back,

Thx much for the testing bro, I for one appreciate ^^
http://images1.wikia.nocookie.net/ff...1H_1.0-2.4.jpg
This is Masamunai's solid testing on this. This is where some of the values on wiki are coming from. I'd trust this right now before I'd trust the wiki, to be honest (although it is incomplete). This is for 1-handers only.

pDif cap, crit hits max is 3.15. Min is ~2.65. Divide max/min, u get ~1.189. This is around the spread that you want to try to reach if you want a 'sufficient sample'. Otherwise one sample in your comparison may be more sufficient than another, and you'll get a very misleading conclusion (or lack of a conclusion) if you try to compare them.

Divide your max/min's. In your non-fatality sample, you get 1.177. For the fatality one, you get 1.190. So your second sample is sufficient. The fatality one is higher than the real pDif max/min. That doesnt concern me since it probably has to do with rounding/truncation. What does concern me is that you said both equip sets had identical dex amounts. This means there is no reason why fatality should have a lower max or min. So hypothetically, your max w/ fatality should be at least 613 (your non fatality sample), bringing it over the 1.189 mark even further, showing a bigger spread of pDif than Masamunai's testing shows. That's its whole other can of beans though.

What's important for now is that your non fatality sample could use more sampling to reach sufficiency. I'd wager that your max wouldn't go (much) higher, but your min would go lower.

Actually scratch that. What's actually important is that even if you reached sufficiency in that sample, it'd still show (from what I can tell, looking at it), that fatality belt is not giving you any improvement at all. Meaning, IMO, one of two things that are most likely happening:
(1)The conditions to activate the latent have not been met, or
(2)The latent is active, but the "increases crit dmg" is actually working like a hidden boost to your attack value. But, since your attack/pDif is already capped, it's not able to increase your damage any further in this situation.

Given that X's does work on level 1 mobs and has the same exact reading on it, I'm wagering that (1) is much more likely :3

So I wonder what the latent is.. hmmz

He didn't do any special conditions either. Just white HP, sa+ws one-shot killing level 1 mobs, same stats w/ and w/out belt. One of his samples is sufficient and one is not. What bothers me, as I mentioned above, is that based on his actual dmg done (assuming he didn't typo) if you combine his two samples, his spread of damage is more than what our pDif information tells us is actually possible, and it appears that the latent isn't active as well.


tldr: These samples are not yet sufficient. Most of them can't yet show if the latent is active or not active. Go beat on more mobs!

[Raelia]

The problem you all face is multifold.

One, you're using a one-handed weapon, meaning your pDIFa will be below 3.0 (uncapped) 50% of the time at capped ratio. This is addressed below.

Two, you're not hitting hard enough to really be able to pick out a percentage gain. With the numbers in the OP, a 5% gain is completely indistinguishable from the spread (10% is available on X's Knife or Cruadin, I don't expect the Belt to exceed this).

Something else entirely is that you've utterly miscalculated your pDIFb min and max, using parroted incorrect numbers from parses that were actually insufficient to pick out the true 'minimum' (which is fairly unreachable after secondary randomization). I strongly suggest you read this.

Your maximum pre-crit pDIFa is going to be 2.0 * 1.2, and minimum is 2.0 * 0.8, 1.6. These multipliers do not differ from 2-handers, only the Ratio cap is different, but the Ratio cap is the kicker when it comes to testing with crits because a 2-hander is far more consistent.

Now that we have a proper pDIFa range, 1.6-2.4, we can add 1.0 for the critical hit. Our range is now 2.6-3.4. The 3.0 pDIF cap (the only pDIF cap, everything else is a mathematical limit) splits this right down the middle. You have a plain 50/50 chance of seeing a capped crit, so half of your data is potentially worthless.

Two handers, on the other hand, wind up with a crit pDIFa range of 2.76-3.64. Fully 72% of 2-hand crits reach the 3.0 cap.

If you don't know what I'm talking about between 'pDIFa' and 'pDIFb', you definitely need to scroll back up and click that link. It is very relevant to why 'capped pDIF' crits still vary between 3.0 and 3.15.

Each of these is then subject to the 1.00-1.05 secondary randomizer. Your minimum is not affected (so as to be the 'minimum', but in practice you're unlikely to ever see it, as this secondary randomizer always multiplies up), and your maximum 3.0 after capping becomes 3.15.

Cruadin and X's Knife function as a 10% increase in crit damage, similar to '+% WS damage' and other final damage modifiers. When I tested Cruadin's 10% crit bonus, it was fairly easy. An 83 damage weapon with capped +17 fSTR is 100 damage. This made it rather painless to simply read my pDIFb right out of the log. 3.465 was the expected pDIFb maximum for a 10% bonus, and after several 346 damage criticals (Tags on, killing worms in N.Gusta(S)) I was more than satisfied with my result. I'm never interested in finding averages, which I liken to trying to put an IKEA chair together with the lights off.

All of this is irrelevant though. Its just my tendency to bash knowledge into people. What you need to be looking for to find the latent, then, is a critical above 3.15 * Base Damage.

Beast Slayer, DMG 42 WRANK 4 fSTR cap +12, 54 base damage.

54*3.15 = 170.1. That ".1" is going to be floored to 170 (like my 346.5 above), exactly what the OP hit just once and what Shamaya already covered.

If the bonus is 5%, you will see up to 178(.6) damage. If 10%, 187(.1) damage. The decimals are important, as the smaller they are the less likely you'll hit that fractional range to see that number pop up, and instead most likely see 1 damage point less.

Stop all this shite about sufficient sample sizes, average pDIF, and wonky irrelevant numbers that make me 'Waht is this I don't even'. OP has simply not found the latent. Masamunai's numbers are spectacular, but don't provide anything but a hindrance to people actually learning the new damage calculation (he provides just the numbers, I provide the math that gets those numbers), see my link above or the pDIF page on the wiki, which was written by me and works.

[miokomioko]

I'd like to believe that the game would not be able to judge when the "kill shot" would land however.

I'm pretty sure it knows whether a Stormwaker automaton's spell will kill a mob or not and uses the lowest-tier spell that will kill... It makes me uncomfortable.

[Psion]

Still waiting for anyone at all to try testing during the full or new moon, or during the vanadiel months of june/july. Despite the name of fatality, it may very well be related to the name of the scorpion that drops it.

I know their pain in the butt variables to test for but it's still something that's quite possibly the latent.

[Shamaya]

Stop all this shite about sufficient sample sizes, average pDIF, and wonky irrelevant numbers that make me 'Waht is this I don't even'.

I think your post is a good contribution. Personally I've read it, but I couldn't bear to read the whole thread, it got kinda messy after awhile. Actually you pointed out a good thing that I didn't think of before (the 50/50 thing due to the floored 3.0 cap).

Three separate questions,
(#1)How do I know if the latent is active?
(#2)How do I know if, assuming I continue to increase the sample size, I might or might not find out if the latent is active?
(#3)How do I know the potency of the latent's effect (dmg+%)?

These are all good questions, #1 being the foremost concern. And I agree, you should,

What you need to be looking for to find the latent, then, is a critical above 3.15 * Base Damage.

When this happens, you will know that the latent is active.

If it does happen, then proceed to #3. If it doesn't, check #2.
Only thing irritating about this for the lamen tester is that they won't be able to look at a number and tell if it's above 3.15. To those guys, I'll tell ya what you need to be looking for. You're looking for bDmg + fStr: [(your weapon's base damage) + (that base damage, divided by nine; remove the decimal) + 8] * 3.15

IE for a 39 base damage, it becomes 39 + 4 + 8, or 51. Then multiply by 3.15. That's what your maximum possible dmg should be
Assuming a level 1 target (no piercing/slashing/blunt affiliation change). And also assuming that your strikes can crit, and that you're not Thf main doing sneak attacks.


#2 is about sufficiency. It's discouraging for lamen testers to go out thinking the community wants them to kill 100+ bunnies, but that's not so. There's an easy way to know if you will gain from increasing your sample. Simply divide pDifMax/pDifMin (in this case, 1h crit cap). The number you're looking for is ~1.189.
(a)If your sample is close to this number and you haven't seen a dmg sample higher than the normal (as Raelia describes in #1), then you can be sure the latent isn't active; at least, in all probability.
(b)If your sample isn't close to this number, then the latent may be active, but you may have just not hit high enough to see it yet. So you can't be sure yet. However, given what Raelia said, if the latent is active then it's likely you'll see it in your sample pretty soon; I'd guess very likely within a dozen or few kills.


Regarding #3, The only way I can see for finding out what the "enhances whatever-whatever effect" is doing, is by first making sure your sample is sufficient, then comparing (i)your numbers to, (ii)the numbers you'd expect, assuming the latent wasn't active. If you find that your latent is active, continue the sample until the max/min is ~1.189. Then, either of the following will work to allow you to find the potency number: (a)divide max of latent active, by max of latent inactive, (b)do this, but for min instead of max, or (c)do this by dividing averages.
For example, if your highest dmg done is 400 and lowest is 200 (just pretend) w/ latent active, whereas max is 200 and min 100 with latent inactive, 400 / 200 = 2. 200 / 100 = 2, and 300 / 150 = 2. You get the same answer any way you cut it. So the latent would be doing 2x (aka +100%) dmg.

[CDF]

This has no bearing on latent discovery, but for one-handed weaopns, if pDIFa minimum is 2.0*.8 = 1.6 and the random factor only increases pDIFa (or pDIFb), why is the observed pDIFa minimum closer to 1.55 (looking at talk discussion on FFXIclopedia)?