Blue Mage: Best thread ever

[Prothescar]

double

[dasva]

Well that's only because you did /dnc. If you were /sam it would've been cake. Delicious turtle cake

[Prothescar]

Would have taken much, much longer too waiting on Sekkanoki. 8.3 hours to be specific.

[VZX01]

Hmm so you'd sleep the mob, perform step->RF repeatedly as necessary to reach 100 TP (or not? because I found subtracting from random TP amount when meleeing could be more annoying), engage while spamming the /ws "Chant Du Cygne" <t> macro so you don't get a single swing before WS unleashed?

Hopefully this is the right thread, but I believe it was brought up a few pages ago: BLU tanking Hahava. I just solo tanked it on BLU/WAR. Never got into red hp the entire fight, except once when the weakness tp move went off. It was actually extremely easy, and much easier than crab guy. I went full out defense, taco, cocoon, defender, and obviously my damage totally sucked. I didn't even bother WSing. All I did for hate was spam fantod, actinic burst, animated wail, and magic fruit after his spells went off. None of the mages died the entire fight, had 3 BLMs, 2 SMNs.

Overall, not as efficient as PLD, but definitely works very well. Perhaps even easier than PLD, just not as fast. You technically don't even need to melee and can just stand there spamming fantod (no tp feed).

So you sit at 300TP almost in entire fight? Why?

[Prothescar]

No need to sleep them. No need to spam either, simple enough to time it.

[dasva]

Timing the loss of 200 tp while mob is awake? Now I question your method

[Prothescar]

Not entirely sure what you mean by "timing the loss of 200 TP". It isn't difficult to surmise how to do this, it's quite simple really.

[Byrthnoth]

OverTPing does not invalidate the test, so he doesn't need 200 exactly every time.

[Prothescar]

I never had more than 15 TP over the value that I was testing, normally about 5-8, which would result at most to +1% to crit rate. If it was over 10 I kept the number in a separate pool of results to be safe, and was the result of an attack right before the ws.

[dasva]

Yeah I guess. Just a little less percise but you weren't really going for exact values so guess it works

[VZX01]

Then the numbers above should have a ±1% crit hit rate error margin.

[Byrthnoth]

Naw, the exact "can't be wrong" interpretation of the data is the way I phrased it. The test is effectively one sided.

Increasing sample size increases how sure you are that the current crit rate is actually 100%. I'm not sure exactly how real statisticians would do this, but the basic idea would probably boil down to:
.99^n = odds of n consecutive crits if your crit rate is actually 99% instead of 100%

[Prothescar]

While not impossible, when you get down to a 0.66%~2% chance of not seeing a non crit and you don't see that non crit, it's a fairly accurate test. I'm certainly not saying these are exact to-the-decimal values, but they're close enough.

[Helel]

Hmm so you'd sleep the mob, perform step->RF repeatedly as necessary to reach 100 TP (or not? because I found subtracting from random TP amount when meleeing could be more annoying), engage while spamming the /ws "Chant Du Cygne" <t> macro so you don't get a single swing before WS unleashed?



So you sit at 300TP almost in entire fight? Why?

Like I said, my damage totally sucked. The most my WS did was 80 damage. This was our first run. I didn't know what to expect so I went full out -pdt (with -pdt sword and genbu's shield). In terms of hate, it was more efficient to spam fantod than to WS.

[dasva]

Um pretty sure he was responding to the CDC testing...

[Prothescar]

The 300TP bit was in response to hahava tanking.

[Byrthnoth]

Just for future reference, I wrote up my understanding of the 100% critical hit method here.

[dasva]

Damn quoting the whole thing!!! And either way throw out a vorpal blade or something geez

[VZX01]

Just for future reference, I wrote up my understanding of the 100% critical hit method here.

Statistics was a 7-year-old story for me, so I'm not gonna comment on the concept.

IMO, It'll be easier to read (and understand) if you replace every "non-crit" word to "regular hit".
And... ultimately:

At the final level, you should end up not seeing any non-crits and Critical Hit 303 times in a row,

will be easier to read if it's changed to

At the final level, you should end up seeing nothing but 303 critical hits in a row,

Because double negative is an extra processing overhead in the brain (at least for me)

...Just a suggestion...

[Alistaire]

You see, you'd think that the test server would prevent idiotic, heavy handed nerfs like these. Was that not the goal behind it, to have the players voice feedback on stupid decisions like this? What's the point if they're going to ignore it all anyway?

Making this my new sig as this cannot be repeated enough.

Edit: Oh, and to answer the question the point is to save SE money on paying their people to test things.