Fin sushi whats the deal?

[Tomk]

Mele ACC and +INT? Sounds like it's perfect for elemental weapons skills since they are fint instead of fstr. I wonder if that is what SE had in mind?

[Wizerd]

There is no significant Magic Accuracy on Fin Sushi. After only 83 casts, it wasn't worth my time to keep testing. My results fell within 5% of my base testing (which has a sample size of 500), when it should have fallen well outside that if it had a percent increase. If it does have any Magic Accuracy, it's something pitiful like +5, and not worth eating the food.

May or may not finish testing later, I dunno. G'night~

[Callisto]

Thanks for testing a bit at least...I personally still use this when dicking around as melee RDM for things like Diorite/Waters farming, wouldn't be terrible for Nyzul either(at least in our specific setup once I've lined up the next pull or two I usually melee to speed things up).

Edit: Not sure if anyone noted, but this also does not have the DEX like most Sushis, only the INT. Still, as RDM I'd take Acc + 5 INT over Acc + 6 DEX/5 AGI and whatnot.

[Halon50]

Thanks for the tests!

[ronin sparthos]

Now to test Octopus Sushi... something is telling me that it has 1% Double Attack rate like Tentacle Sushi but doesnt have the cap on it.

[Maxxthepenguin]

There is no significant Magic Accuracy on Fin Sushi. After only 83 casts, it wasn't worth my time to keep testing. My results fell within 5% of my base testing (which has a sample size of 500), when it should have fallen well outside that if it had a percent increase. If it does have any Magic Accuracy, it's something pitiful like +5, and not worth eating the food.

May or may not finish testing later, I dunno. G'night~

...you went through that 500 casts to establish a base and only did 83 casts to figure out fin sushi?

You still didn't prove anything...

[Wizerd]

...you went through that 500 casts to establish a base and only did 83 casts to figure out fin sushi?

You still didn't prove anything...

I didn't prove that it didn't have any Magic Accuracy on it, no. I did, however, prove that it did not have +10% or +15% Magic Accuracy on it, the latter being what might be expected based on melee sushis. When I tested NQ staff, the results shot up from 56% to 82% (or something like that) accuracy - a very noticable difference. Something even 100 casts could reasonably prove. Theoretically, if Fin Sushi had +15% Magic Accuracy on it, it should have brought me to 96% Accuracy (or 95%, the cap for full hits).

Also, I said it wasn't worth my time to continue testing, because to me, a food with low Magic Accuracy on it is not worth the difficulty or price of acquiring it (Kalkanbaligi, as far as I know, is not an easy or super common catch).

In case I didn't make it clear, the 500 cast base was already set from other Magic Accuracy testing. I didn't do 583 casts to test the Fin Sushi, I did 83 under the same controls as the 500 casts, and compared results.

I still have 7 pieces of Fin Sushi left, and I have the 83 cast results recorded, so I may continue later just because. But for now, I'd rather use my time meriting or some such, or figuring out exactly what skill / INT does for Magic Accuracy.

[Midnightjade]

Did anyone check crab sushi yet? Wiki says it doesn't give any acc, but I find that very hard to believe. It also seems to be selling like crazy on my server.

If nobody's tested it, I guess I could give it a go - should also be able to make it in a couple of days (currently 24k/30k on the guild point requirement).

Info welcome, and thanks in advance. And yes, I did try a search before asking.

[tymora]

There is no significant Magic Accuracy on Fin Sushi. After only 83 casts, it wasn't worth my time to keep testing. My results fell within 5% of my base testing (which has a sample size of 500), when it should have fallen well outside that if it had a percent increase. If it does have any Magic Accuracy, it's something pitiful like +5, and not worth eating the food.

May or may not finish testing later, I dunno. G'night~

Be nice if you included what spells and on what day you cast them on. If not using Obi and depending on spell, day can have some random effect on spells which can skew results.

There is also a mob level variation to take note of.

[meson]

I didn't prove that it didn't have any Magic Accuracy on it, no. I did, however, prove that it did not have +10% or +15% Magic Accuracy on it, the latter being what might be expected based on melee sushis. When I tested NQ staff, the results shot up from 56% to 82% (or something like that) accuracy - a very noticable difference. Something even 100 casts could reasonably prove. Theoretically, if Fin Sushi had +15% Magic Accuracy on it, it should have brought me to 96% Accuracy (or 95%, the cap for full hits).

Also, I said it wasn't worth my time to continue testing, because to me, a food with low Magic Accuracy on it is not worth the difficulty or price of acquiring it (Kalkanbaligi, as far as I know, is not an easy or super common catch).

In case I didn't make it clear, the 500 cast base was already set from other Magic Accuracy testing. I didn't do 583 casts to test the Fin Sushi, I did 83 under the same controls as the 500 casts, and compared results.

I still have 7 pieces of Fin Sushi left, and I have the 83 cast results recorded, so I may continue later just because. But for now, I'd rather use my time meriting or some such, or figuring out exactly what skill / INT does for Magic Accuracy.

SE did just say they were thinking about revealing certain food's hidden effects. This to me says there's some foods out there that we aren't utilizing that we can be. Could also mean they want us to know about the uber +def food that we've been ignoring.

That being said, you proved that there isn't an uncapped Magic Accuracy effect which I think everyone would be very surprised if there was one. But even if its capped at say +10 Magic accuracy, it would be nice to know.
You may have posted about your previous tests and listed what mobs,spells, skill levels, gears, and stats you were using. Can you either list them again or link that thread? If anyone else know the link can they list it?

[evilpaul]

Now to test Octopus Sushi... something is telling me that it has 1% Double Attack rate like Tentacle Sushi but doesnt have the cap on it.

Octopus was STR+1 and the usually Acc+% with a cap no lower than 45 when I tested it a while back. I don't know whether it gives DA or not I always had a Brutal Earring on with it.

[Wizerd]

I didn't realize this thread got bumped haha.

Tymora: I tested using Water, all casts were done in a single day, and the day was neutral to water. I forget what day now, but it was definitely not Watersday or Lightningday. It may have been Earthsday, to the best of my recollection. There was no level difference to take into account, as I tested on Even Match Qiqirn Archaeologists.

Meson: I'd be glad to list what I used.

Area - Aydeewa Subterrane
Monster - Qiqirn Archaeologist (Even Match only)

Spell used - Water (tier 1)
Elemental Skill - 238
INT - 66 (Base INT is 66, obtained 66 INT by using Custom Vest and Devotee's Mitts, for a total of -5 INT while the food was in effect. The food itself gives +5 INT, which needed to be offset properly.)

Otherwise, I was naked.

Previous test results, without food and with identical stats to the above:

Water - No staff, neutral days only
400 casts (225, 110, 34, 31)
56.25% full landing rate

Fire - No staff, neutral days only
300 casts (168, 84, 26, 22)
56.00% full landing rate

If one is to assume that the above two spells have the same base magic accuracy:
Combined Water and Fire - No staff, neutral days only
700 casts (393, 194, 60, 53)
56.14% full landing rate

Blizzard - No staff, neutral days only, +12 Magic Accuracy from merits
200 casts (147, 36, 10, 7)
73.50% full landing rate

Water - NQ staff equipped, neutral days only
300 casts (248, 39, 11, 2)
82.67% full landing rate

Water - No staff, neutral days only, Fin Sushi in effect
83 casts (52, 18, 4, 9)
62.65% full landing rate


Keep in mind that the main purpose of my tests was to see if there was a significant bonus, like the NQ staff gave. While testing Water and Fire originally, I had higher returns like the Fin Sushi test, around 62-63%. A couple hundred tests later and it averaged out to a lower number. I forget what the actual damage of the water spells did, I believe it was somewhere around 24, 12, 6, 2?

Also, throughout my testing, I never saw a single 1/16 resist. I'm still waiting for someone to prove me wrong, but I don't think 1/16 resist exists. Capping -magic damage taken can make you feel like you resisted something to that fraction.

[tymora]

Fin Sushi
Sliced, raw kalkanba served on a small
portion of sweet vinegar rice--a
popular dish in Far Eastern nations.

Fin Sushi +1
When the fins taste this good,
who needs the rest of the fish?

Anyone find the item description for +1 funny?

[tymora]

Water - No staff, neutral days only
400 casts (225, 110, 34, 31)
56.25% full landing rate

Water - NQ staff equipped, neutral days only
300 casts (248, 39, 11, 2)
82.67% full landing rate

Water - No staff, neutral days only, Fin Sushi in effect
83 casts (52, 18, 4, 9)
62.65% full landing rate

Though 83 casts with Fin sushi is a little low, we're looking at ~ +5% macc? That would put it at around +10 macc at 75?

Be nice to see if having NQ staff + sushi would boost to ~87.

[Sylphet]

For the record, I'm a PhD biologist so yes I do know what I'm talking about.

Just a quick lesson in statistics. Since I very commonly hear people say things like "X number is not enough to say anything.", I'd like to show you a way to deal with data sets so that with some manipulation, they can say something significant.

Error bars are your friend. Everyone remember standard deviation from high school?

If you have a data set of 100, you can either call it a single set or you can break it apart into several different sets (of varying sizes- e.g. 4x25, 10x10). Saying that you have an average of x over 100 samples does not tell us as much as saying x +/-y. For instance, if you have an average of 50 (scale of 1-100), 50 +/-5 is not the same as 50 +/-25.

To do this you would need to take a set of (e.g.) 25 casts and for each of those obtain a mean. Then you determine the standard deviation of those means. Any basic spreadsheet program like Excel can do this for you.

This is then repeated for your second set of conditions. Then you can compare the statistical significance of the two sets despite the variation in sample size. The simplest way is to look at the error bars to see if they overlap. If they do that, means that there is a possibility that these 2 results are were obtained by chance. A more rigorous way is to obtain a p-value (which is a measure of statistical significance) to determine what the exact chances are that these results were from the null hypothesis (random chance).

Obviously increasing the sample size leads to smaller error bars and subsequently a more accurate assessment of whatever you're testing; however it is not sound when people automatically dismiss results from testing because of the sample size.

[Wizerd]

Though 83 casts with Fin sushi is a little low, we're looking at ~ +5% macc? That would put it at around +10 macc at 75?

Be nice to see if having NQ staff + sushi would boost to ~87.

No, we're not looking at +5% M.Acc.

If we were to assume that all of my results had little margin of error (clearly not the case), we could assume that 1 Magic Accuracy increases your magic hit rate (based on full hits only) by 1% (which I do believe is the case, despite a decent margin of error). Both the Blizzard tests with +12 Magic Accuracy and my test with NQ staves (assuming Magic Accuracy is calculated similar to Melee Accuracy) fall relatively in line with this assumption.

Based on that, the difference between the ~62-63% hit rate with Fin Sushi and ~56% hit rate without it would mean that Fin Sushi has a total Magic Accuracy of +6 or +7. Of course, since there were only 83 casts, the margin of error is high enough that the possibility of there being 0 magic accuracy on the Sushi is quite decent. By the same token, it could also have +10 or +12 Magic Accuracy, if someone wanted to take the extra time to test it.

What I concluded was that the sushi did not have anywhere near +10% Magic Accuracy, and therefore was not worth using based on the supposed difficulty of catching the fish. After 83 casts, I felt confident enough to deem the sushi worthless, and I clearly stated that I did not prove that it had nothing on it. I would wager that a total of 400 casts would be sufficient to prove whether or not it had some magic accuracy on it. I don't know what the margin of error would be for that sample size, but I'm sure Sylphet could tell you. I don't know.

PS: +5% Magic Accuracy should put you somewhere around 69% full hit rate, give or take a couple percent.

[Tordall]

Since people seem to care about error propagation (or uncertainty):

Water test w/o sushi had a rate of .56 +- .025, water test w/ sushi had a rate of .63 +- .053. The difference between them was .07 +- .078. So, the difference isn't really meaningful (since it's consistent with no difference).

[tymora]

Spamming Fortifications with spells seems a good way to do spell testing. Lasts forever, high defense, would think high resist rate too.