Magic Accuracy and INT

[pchan]

We do know certain HNMs are highly resistant to enfeebling magic but we don't really understand the mechanism behind.

We do. If you resist a lot, your magic accuracy is not high enough, that IS how it works. Otherwise, why would SE even let you land threanodies at all ? You don't seem to understand the concept. Your hit rate is necessarily between 5% and 95% (except immunities / change of "cap" values which seem to happen with resist - traits). 5% means you have -90 magic accuracy than the mob or less, 95% means you have 45 more magic accuracy than their magic evasion, or more. Simple. Let's say the max magic accuracy a RDM can get is 400. It's easy for SE to gimp your ability to slow the mob while not making it immune : give him 500 magic evasion. It's easy for players to resist monsters spells : up your own magic evasion with (-> fire resists for instance). You are not happy that Tiamat systematically hurts you with firaga III ? All you need is 90 more fire resist than its magic accuracy. Game mechanics apply to HnMs and we do understand it.

In the same trend : clearly it's easy for SE to make puddings very resistant to silence, flayers very resistant to poison. This is not "special" mechanics.

[Shinryuu]

Yeah I agree with pchan here, it's not special mechanics to put say, 300 extra resist wind on a pudding or +magic evasion for silence or whatever. Only 'special' mechanic is HNMs building resistance to something.

[Uzor]

Having played with Pchan for a few years, I would not be surprised in the least if he's crazy enough to actually do a test like this manually. He's done it before. Whenever I logged on I used to see him in, say, the underwater ruins doing tests on rats, or doing tests on puddings, or doing tests on puks etc. etc...

I never understood how he had the patience to do it. I'd understand if he botted, but unless his bot is sophisticated enough to talk while performing his tests... I mean, fuck, I barely have enough patience to exp or merit for 2 hours.

[Nevex]

You are not happy that Tiamat systematically hurts you with firaga III ? All you need is 90 more fire resist than its magic accuracy.

Mobs' mechanics =/= players' mechanics.

Gained TP from being hit is a clear example, I don't see why m.acc/m.eva should be any different.

Even if the +90 m.eva/elemental resist is true for Tiamat's Firaga III, it might be entirely false for something else. Since each spell/TP move has its own base accuracy, you can't generalize.
Flash, Stun and Dispel are clearly more accurate than most of the other spells, so is Anticans' Jamming Wave TP move just to name a few cases.

[pchan]

Mobs' mechanics =/= players' mechanics.

Gained TP from being hit is a clear example, I don't see why m.acc/m.eva should be any different.

Even if the +90 m.eva/elemental resist is true for Tiamat's Firaga III, it might be entirely false for something else. Since each spell/TP move has its base accuracy, you can't generalize.
Flash, Stun and Dispel are clearly more accurate than most of the other spells, so is Anticans' Jamming Wave TP move just to name a few cases.

It will be true for any magic spell as long as you counter it with approriate barspells and the magic accuracy is not to high. I guess if you stacked light resists, light carol etc you would resist it decently. SE is doing their mobs. If they want DL to be stun resistant they do it. If they want omega not to be stun resistant, they do so (nin/drk skill sucks, they have like ... 120 magic acc ?).

[Demosthenes11]

lol @ being a butt hurt bitch when someone asks you methodology. when you post data sets, you better be prepared to answer questions, otherwise don't even fucking post it.

your stupidity is incredible in this regard and quite frankly I don't see why not bannable.

[pchan]

lol @ being a butt hurt bitch when someone asks you methodology. when you post data sets, you better be prepared to answer questions, otherwise don't even fucking post it.

your stupidity is incredible in this regard and quite frankly I don't see why not bannable.

pay 50$

[Callisto]

The reason for this is that Cerb + Khim (and others) build up resistance to certain enfeebles. Landing the first slow/elegy/para is not as bad as trying to do it halfway through the fight. It's sort of like the stun resist that Baha or Khim build up, or like the grav resistance many NMs build up (and then slowly fade.)

While I admit that my HNM experience is pretty limited since I've never focused on that content, I'm still iffy on this. The only instance where I've had a noticeably easier time landing debuffs at the beginning of a high-end fight than further in has been due to me starting with TP and tossing out a Death Blossom to open. I would attribute it more to a placebo that you just happen to run into a longer chain if resists halfway through since your actual land % is relatively low still.

Of course Gravity, Sleep, and Bind would be exceptions, I know many NMs specifically build resistance to those spells, but I've never seen the same thing with the likes of Slow/Para/Blind/Poison/etc.

[Tagus]

lol @ being a butt hurt bitch when someone asks you methodology. when you post data sets, you better be prepared to answer questions, otherwise don't even fucking post it.

your stupidity is incredible in this regard and quite frankly I don't see why not bannable.

I agree with this 100%. You have provided no proof that this test was actually conducted, and to think that you would spam bind 44,000 times over the course of 50+ days without documenting it is almost dumber than making up the data in the first place. At least have somebody, a friend, a LS mate, somebody with a reputation that isn't laughable, come in here and say that they saw you doing tests, heard you complain about it, had you skip an event for testing... something more than "I don't have to prove shit."

[Demosthenes11]

I agree with this 100%. You have provided no proof that this test was actually conducted, and to think that you would spam bind 44,000 times over the course of 50+ days without documenting it is almost dumber than making up the data in the first place. At least have somebody, a friend, a LS mate, somebody with a reputation that isn't laughable, come in here and say that they saw you doing tests, heard you complain about it, had you skip an event for testing... something more than "I don't have to prove shit."

that's not what I have a problem with. Hell, I don't doubt he actually did it. When you post scientific research, however, you have to be prepared to take criticism, doubts, and answer questions relating to the experiment. That's why it's science. If Pchan's line of thinking prevailed, nobody would post screen shots to prove anything and AV would have been "killed" 50 times.

[Tagus]

that's not what I have a problem with. Hell, I don't doubt he actually did it. When you post scientific research, however, you have to be prepared to take criticism, doubts, and answer questions relating to the experiment. That's why it's science. If Pchan's line of thinking prevailed, nobody would post screen shots to prove anything and AV would have been "killed" 50 times.

True... and a major part of preparing to take criticism when presenting data is following the scientific method, which includes documentation throughout the test. With this half-assed bullshit he's throwing at us, the only way it can ever be verified is through a complete reproduction. What I'm saying is, if he actually did the test (which I personally find questionable), it might as well not have even happened if he's not prepared to explain how he did it or prepared to prove that it was done at all. I understand when people say "lol scientific method video game etc", but he should have expected criticism.

The ffxiah profile pic was a nice touch though.

[rog]

How exactly do you expect him to prove he did it? Screenshots? That doesn't prove anything, he could go run to sea now bind a hpemde, and ss that. Explaining how he botted it? That doesn't do it either, i could explain 10 different ways to bot this test right now, i'm sure he could find at least one way to explain that. Having someone else say he did it? I'm sure he could find at least 1-2 people to say he did something he didn't do, so again, not proof. Copy of chat logs? Well, that would be significantly harder to fake, but could certainly be done.

[Tagus]

How exactly do you expect him to prove he did it? Screenshots? That doesn't prove anything, he could go run to sea now bind a hpemde, and ss that. Explaining how he botted it? That doesn't do it either, i could explain 10 different ways to bot this test right now, i'm sure he could find at least one way to explain that. Having someone else say he did it? I'm sure he could find at least 1-2 people to say he did something he didn't do, so again, not proof. Copy of chat logs? Well, that would be significantly harder to fake, but could certainly be done.

Well, to echo a statement that has already been made with no response from him of any kind, maybe he could provide some form of media to show us he did the work.

[Yugl]

Alright guys, lets settle in for over 10hrs of half-nakedchan spamming bind.

I skimmed that quick and I saw 10 hours of half naked pchan and my eyes glazed over for a minute. Damn you Yugl.

Yeah, media that Sath can enjoy LOL.

Edit: I hope this is becoming clear for all of you demanding media. Pictures can be edited or dated from any time. A video would be unrealistic. Parse data can be edited. What else is there? The entire purpose of asking for methodology is so that it can be repeated and the results compared. Media has nothing to do with proving the legitimacy of the test.

[Byrthnoth]

I hate to do this, but it's the easiest way to stop the bickering/bitching and get some productive conversations going in this thread.

This is what started the whole "maybe pchan faked his data thing:"

Of course there is always a chance, but running ~7500 trials (trial1) and hitting 49.99% is crazy. But let's just give that the benefit of the doubt. Weirder things have happened. Tests 4/5 (the ones that conflict with Lodeguy's data and are most relevant) hit what are essentially the most perfect value you can want given your ultimate prediction of dINT=15 is the cutoff, not dINT=10. The absolute perfect value for test 4 is T1Average + 18, or 68%. You got 68.6%. The absolute perfect value for test 5 is T4Average + 2.5, or 71.1%. You got 71.3%. I know the error bars are tight due to the large trial count, but the phrase "too perfect" comes to mind. We all know the "law of large numbers" tells us that eventually we will hit the true average with infinite trials, but it just sets off red flags for me.

You can't look at his data and prove that he faked it, nor does the data imply that he faked it in any way shape or form. If anything does, then what implies that he faked is the amount of effort required for this testing method and how minimal the finding ever had the potential to be, but he might really just be crazy for all we know. Still, you can't use "your data looks too perfect" as an argument. It isn't like he tossed up perfect values.

If I'm not mistaken, ideal values for his model (not including the flooring he suggests) would be:
T1: dINT = -3: 50% :: 49.99%@[49.02,50.98] Obtained -- 0.02 standard deviations away
T2: dINT = +40: 80.5% :: 80.33%@[79.24,81.48] Obtained -- .2975 standard deviations away
T3: dINT = +10: 63% :: 62.49%@[61.24,63.79] Obtained -- .784 standard deviations away
T4: dINT = +15: 68% :: 68.62%@[67.32,69.98] Obtained -- .914 standard deviations away
T5: dINT = +20: 70.5% :: 71.38%@[70.15,72.66] Obtained -- 1.37 standard deviations away

Less than half of his values are inside the 50% confidence interval (.675 standard deviations away). 80% of his calculations are inside the 80% confidence interval. All of his calculations are within the 90% confidence interval. This doesn't scream faked data to me.

As a side note:
He suggests the flooring might occur, but I don't think that's likely based on the T5 result. If it does occur, then the boundary is at least dINT = 16.

[Tagus]

I actually think the easiest way for the bickering to end would be for someone to accompany PChan on another trial to verify the test's validity. I might have been able to convince somebody to do that, if only it was still 2006...

[rog]

I actually think the easiest way for the bickering to end would be for someone to accompany PChan on another trial to verify the test's validity. I might have been able to convince somebody to do that, if only it was still 2006...

One night's worth of botting = proof that all the rest were done?

Not saying he faked it, but seriously, i don't see how you expect him to prove he actually did it.

[Tagus]

One night's worth of botting = proof that all the rest were done?

Not saying he faked it, but seriously, i don't see how you expect him to prove he actually did it.

Is it so wrong to expect him to try?

[rog]

Is it so wrong to expect him to try?

It is when you expect him to do something that proves nothing.

[Tagus]

It is when you expect him to do something that proves nothing.

Oh, so we're talking expectations now. Well, for the record, my expectation is that he will provide us with absolutely no evidence that this test actually occurred, not because he doesn't give a shit, but because the test didn't occur. Mostly I'm expecting this because I know that if I did an obscene amount of work and people called me a liar, I'd get pissed off and try to prove them wrong. Unless of course, I didn't do the test, in which case I would say, "I don't owe you shit".

But maybe he thinks differently than me due to the abuse he takes here and truly doesn't give a shit, so I'm just asking for proof.