Random Facts Thread: Other

[Byrthnoth]

https://www.bg-wiki.com/bg/Magic_Hit_Rate

I put CDF's compilation of Magic Hit Rate analysis on the Magic Hit Rate page because that's what was actually being measured. MAcc and MEvasion aren't values we have any way to quantify except as they relate to magic hit rate.

[pchan]

The C skill for player's magic evasion is pulled of thin air, and was present before I edited the pages. I've conducted plenty of tests against qq poulterer, imps and flan, but they are no longer available to download.

The results from the wiki that I have tested are

int <-> mac relation for nukes and sleep.
hit rate <-> mac (from which I deduce a meva check) for nukes and sleep
skill <-> mac for nuke and sleep
upper and lower caps of hit rate for nukes and sleep
partial resist rates of nukes and enfeebling
silence and paralyze partial resist states
bar-element <-> hitrate relationship (+1 barelement = -1 macc)
HQ staves=30 macc
I remember doing a death blossom test too that had it between -10 and -20 macc

The stuff that is assumed but untested are

-the actual relation "magic evasion" (found on gear) to hit rate is untested due to the lack of gear with such stats. It is an assumption that the meva stats found on gear equals what I call magic eva on wiki.
-the level correction effect on macc is a pure assumption, but a level 40 char can stun a level 75 one so..
-The effect of barstatus effect is pure assumption (eg the fact that basleep affects sleep spell's meva )
-mnd vs mnd check for para or slow is pure assumption.
- higher tier nukes being more accurate that lower tier nukes is from SE's words, it's not tested or quantified (I think a JP that was quoted from cdf is his blog did it, but unsure).

[Yugl]

The level correction sounds reasonable (Untested, but reasonable). Stun and Flash have absurd land rates against non-SDT mobs compared to other spells, so they could simply bypass the level correction component.

[dasva]

Idk they generally even have a good chance to land on stuff that are really resistant to those elements too. I figure either those spells just have absurb base macc based on whatever or most mobs just have low meva to those spells

[pchan]

Mobs have just zero resistance (meva) to stun for most of them. It's easy to check that stun can be naturally resisted through the use of barthunder and elemental resistance gear, get low level char and high level one with stacked barthunder and you will get resists in pvp. The fact that you don't see resist normally (even 5%) would imply that the spell has different levels of resists.

[Byrthnoth]

I would just say stun has multiple resist states and gets a large macc boost. With three resist states, like elemental magic, stun would have only a 0.01% chance of resist at capped macc. We know that there's always a chance (low, but I'd be surprised if it's not ~5%) of Stunning monsters like Qilin that resist but are not immune to Stun (similar to Silence). Things like Adamantoise I don't think can be stunned at all.

Actually, thinking about this again. Something with three resist states would have a minimum land rate of almost 15% (14.2%), which is clearly not the case. I've re-started my Violent Flourish tests to see how this scales.

Stun, flash, and repose all get a large macc boost.

[Ophannus]

Since bst pets can get much higher HP than players (5k+) it wouldn't be farfetched that they might have higher elemental and status resist as well. Wouldn't be surprised if their base attributes, and att/acc/crit were much higher as . Another thing is that something like Monster Gloves/+1/+2 or BST affinity might have a hidden effect of boosting parameters, not unlike wyverns being leveled up. They might get bonus stats even if they're uncapped already but still called with the hands on. Or maybe the Reward III effect gives a boost to m.eva/attribute bonus. Just throwing out conjecture since nobody is certain what exactly Enhances Reward III does over enhances reward I/II.

[fantasticdan]

Uncapped jugs will still summon randomly between 97-99 without relic gloves, 98-99 with nq, always 99 with +2. I doubt there is a hidden effect.

Enhances reward III on relic +2 boots is just +30% healed, compared to +20% on the nq, why they labeled it like that I have no idea.

Enhances reward III on relic +2 body confusion is probably just from nq af body being apparently bugged when it was added. From what I gather based on the discussions I've seen around, it used to be like this:

af body removes: para, poison, blind
af2 body removes: slow, silence, gravity

Than af2+2 was added, which combines the effects and removes all 6, but af1 body starts doing the same thing, possibly the af2+0/1 as well.

[suji]

I picked up Striga Crown the other day and wanted to test its Drain bonus. I did some drains with it on worms EP in Aby LaTheine. Heavens, Ultimate, MM atmas, not Darksday, 390 Dark Magic skill (though I dinged to 391 while testing), no staff. I cast weak nukes to clear stoneskin.

I only have 59 samples so far (a bit more at 392) but I got a couple lows and highs that are interesting so far. Assuming CDF's Drain testing is correct, the lowest I should expect to see at 390/391 skill is 195 and the highest is 391. The observed lowest was 216 (10.7% over min) and my highest was 420 (7.4% over max). I don't think the 216 was a resist since values in that range were more common than base resist rates. So the bonus is at least 7.4% but it seems like it's probably 10%, as the lowest possible value for Drain should be 214 (floor(391/2 * 1.1)).

I'm willing to test more if you guys have ideas on how I could improve. I haven't looked at the Aspir bonus at all yet.

Drain values:

Spoiler: show

216
220
222
223
233
237
245
254
255
258
259
267
274
274
279
280
284
288
288
289
290
294
301
302
302
305
306
309
314
316
318
325
328
332
340
342
349
352
359
364
374
376
377
377
381
384
385
395
397
401
403
407
409
413
413
415
418
419
420

[pchan]

If mobs get MEvasion = 0, the only way to resist Stun would be through "Resist Stun" traits, which would register "Resist! Kirin resists the effects of Stun." in the chat log. That doesn't happen and it's still possible to land Stun on Kirin with a very low success rate. So what you're describing cannot be what happens.

Which means they gave kirin magic some evasion to stun ? Idk. I've fought it 2 or 3 times and landed stun once on it. It probably means they gave him a retarded amount of magic evasion to stun :

if you assume 2 resist states, if your hitrate is at the lower capper you have (0.95)^2=90.25% chance of seeing "resist".
if you assume 3 resist states, 85.73% "resist"
if you assume 4 resist states, 81.45% "resist"


Comparitively if a mob has zero meva, and stun had 2 resist states, you would see a resist 0.25% of the time (1/400) [ 1/8000 with 3 states, 1/160000 with 4 states].


Special magic evasions to mobs already exists. For instance when I tested magic accuracy, I only had 30% landing rate on sleep vs imps while silence, gravity and bind would always land. Similarily puddings resist slience a lot, and bind not at all. Another exemple is flayers that have capped magic evasion against poison.

[pchan]

In the case of kirin, the resistance against stun is probably just a by product of an artifical +barthunder that they gave him, it's not specific to stun. The case of puddings is interesting as it means devs can give mobs element specific magic evasion and also spell specific magic evasion. Which makes me think barsleep does what I think it does.

[Byrthnoth]

So you're proposing:
* We have a base magic evasion stat
* We have Magic Evasion against certain elements (that also apply to statuses), like Fire resist.
** Stun bypasses base magic evasion and magic evasion against certain elements.
* We have Magic Evasion against certain stats, like Barsleepra.
** Kirin, Adamantoise, and other monsters that resist Stun, have something innate like Barstunra that isn't set to 0 (despite the fact that you think Stun in Ballista bypasses Barthundra/etc.?)
* You're implicitly assuming that the minimum magic hit rate can be lower than 5%, because otherwise you can't account for the observed Stun Magic Hit Rate >95% and <5% against different targets.

By the way,
http://www.bluegartr.com/threads/103...=1#post4849940
Yesterday I went back to this and changed my level to 65 (Taru DNC/SAM), then repeated the test (Level 77 Hpemde). I had a 47% Stun rate (~1000 Violent Flourishes). It is obviously difficult to get a high sample size with this due to the method. I had full Dagger skill merits in the first case and none in the second, but otherwise my stats should be as predicted.

Level 65 with no Dagger merits:
1185 Violent Flourishes used
75 Misses (6.3% Miss rate)
1110 Hits (93.7% Hit Rate) - Still consistent with Accuracy +100, base hit rate was 43%
531 Stuns (47.8% Stun Rate +/- 2.9%)


Level 61 with full Dagger merits:
44 Stuns / 448 Hits (9.8% Stun Rate +/- 2.8%)


So I gained 38% Stun Rate with 4 levels. I really believe my original conclusion that I was at the floor and Violent Flourish has 2 resist states anymore. I would need to repeat the test at 60 and still see the same 10% Stun rate to be more convinced. Also, I should note that my margins of error are fairly large. I could be gaining 40% (such a pretty number), who knows!

[Yugl]

Stun and Physical Stun are different though. If you fight the Apademak, Physical stun will not land, but magical stun will land.

[Byrthnoth]

He's also very resistant to Stun. I seem to remember WHM/BLM mules having to use Jupiter's and a MAcc Atma to reliably land Stun on him.

Physical Stun (or Additional Effect: Stun) is less accurate than Stun, but either way we know that Violent Flourish is effected by Magic Accuracy and thus it gives us an automated way to look at Magic Accuracy. We know that Violent Flourish is affected by Magic Accuracy due to testimonies of greatly increased reliability with MAcc Atma, Ascetic's, etc. and I see no real reason to doubt that.

[Yugl]

If the difference is simply magic accuracy, then we have to account for the fact that Draylo could land stun sub DRK, but not sudden lunge on BLU main. We also have to account for the fact that Sudden Lunge has a much greater duration than Temporal Shift, but (I think) Draylo could land Temporal Shift and not Sudden Lunge. The latter is accounted for by saying each type of stun has a different max duration, but how do we account for the former? We're talking about a difference so great that main skill (Blue magic) cannot compete with sub skill level. If sub skill stun is so potent, then we should expect full duration Stun anytime we see full duration Sudden Lunge as BLU/DRK. I'm not so sure that's the case though.

[Byrthnoth]

Temporal Shift, Violent Flourish, Stun, Head Butt, etc. are all the short duration old-Stun. Sudden Lunge is long duration new-Stun. I don't know if he is specifically resistant to that or what (it would seem pretty stupid if he was). Also, Temporal Shift may get a MAcc boost similar to Stun?

At 90, I know that /BLM Stun was not that potent and was not reliable enough to dualbox Apademak NIN+WHM/BLM unless you were using a legit build for it. You can find some discussion between Arthars and other people (I think) from back at that time if you look hard enough.

[pchan]

So you're proposing:
* We have a base magic evasion stat
* We have Magic Evasion against certain elements (that also apply to statuses), like Fire resist.
** Stun bypasses base magic evasion and magic evasion against certain elements.
* We have Magic Evasion against certain stats, like Barsleepra.
** Kirin, Adamantoise, and other monsters that resist Stun, have something innate like Barstunra that isn't set to 0 (despite the fact that you think Stun in Ballista bypasses Barthundra/etc.?)
* You're implicitly assuming that the minimum magic hit rate can be lower than 5%, because otherwise you can't account for the observed Stun Magic Hit Rate >95% and <5% against different targets.

Which mob has a stun magic hit rate > 95% or < 95% ? Magic hit rate only refers to the landing rate of the non resisted spells. If I go on BLM99 stun a wild rabbit I have a 95% magic hit rate, but if I stun serveral hunderd of times I'm unlikely to see "resist" because of partial resist. Similarly I'm offering champaign to anyone showing me 1/8 resist of blizzard V on a wild rabbit.

[Yugl]

If they do, then they have a fairly substantial accuracy boost. At minimum, blue magic reaches 475 (Draylo likely has more knowing him) and sub DRK stun is 150 skill. Any additional magic accuracy affects both of them, so that's null; thus, making the magic accuracy boost substantially exceed 325 magic accuracy. The boost has to substantially beat that because the difference between the two spells is not landing (or very minimal land rate such as 5%) and landing almost full time (One resist during Draylo's fight and Temporal shift landed to cover for that cast).

[Byrthnoth]

Which mob has a stun magic hit rate > 95% or < 95% ? Magic hit rate only refers to the landing rate of the non resisted spells. If I go on BLM99 stun a wild rabbit I have a 95% magic hit rate, but if I stun serveral hunderd of times I'm unlikely to see "resist" because of partial resist. Similarly I'm offering champaign to anyone showing me 1/8 resist of blizzard V on a wild rabbit.

Kirin has a Stun rate of <5% and Wild Rabbits have a Stun rate >95%.

https://www.bg-wiki.com/bg/Resist

With a 5% Hit Rate floor and 95% Hit Rate cap
One State: Minimum observed Stun rate of 5%, maximum observed Stun rate of 95%
Two States: Minimum observed Stun rate of 9.75%, maximum observed Stun rate of 99.75%
Three States: Minimum observed Stun rate of 14.3%, maximum observed Stun rate of 99.9875%

So now you see the problem? The fact that Stun's land rate is >95% indicates it has multiple states using our current resist model. The fact that Stun has a <10% minimum land rate is problematic if you assume a 5% floor with the same model.

[pchan]

The minimal landing rate is verified for elemental, sleep and poison. The visual 10% landing rate for 2-states enfeebles is verified for silence and poison. Idk if there is any data on kirin, but if the model works on him it's a minimum of 9.75% landing rate of stun. I guess one could try to go and stun LBC in salvage (or whichever is stunnable).