Track this topic | Email this topic | Print this topic

A while ago I stated that the redcape graphic associated with the Zephyr Cape, is nothing more than a replacement 'miss' graphic for the +10% MBlock of the Zephyr Cape; nothing more. The redcape subject recently came up again in the Trivia thread so I will post my results from my 2003 redcape study now.

Definitions
RC=Redcape graphic appears followed by the miss text. (a miss graphic which only appears when the Zephyr Cape is equipped)
miss=conical white and red shield graphic followed by the miss text. (ie. the normal miss grahic)
hits=normal physical attacks or special physical attacks.

I also collected information about how often some monsters do their special attacks versus Fight. And I included whether hits or misses occured from a Fight attack or a special attack. I have posted this data as well, for those of you who are interested in the frequency of attacks and specials.

Experiment Group
Here are the test subject's stats:
Locke
2% MBlock (12% with the Zephyr Cape. I intentionally chose Locke because he had the lowest natural MBlock%)
I placed him in the first (of 4) slots. It was a party of 1. He was in the back row.

equipment
No weapon
Heavy Shield
Green Beret
Gaia Gear
Zephyr Cape
Sneak Ring

Leafer and Dark Wind were the test Monsters (chosen because they have all physical attacks).

Control Group
Locke same as above, except -Zephyr Cape and +White Cape (+10% MBlock) to keep the MBlock% the same at 12%.

Results

Control Group
White Cape (No Zephyr Cape)
Total Fight Dives Incisor
Total 300 (100%) 240 (100%) 22 (100%) 38 (100%)
hits 274 (91.3%) 220 (91.6%) 17 (77.2%) 37 (97.3%)
miss 26 (8.6%) 20 (8.3%) 5 (22.7%) 1 (2.6%)

Experiment Group
With Zephyr Cape
Total Fight Dives Incisor
Total 300 (100%) 242 (100%) 22 (100%) 36 (100%)
hits 273 (91%) 222 (91%) 20 (90.9%) 31 (86.1%)
RC+miss 27 (9%) 20 (8.2%) 2 (9%) 5 (13.8%)

* You will need to copy paste these 2 tables into a spreadsheet and set font to Courier to view correctly. This forum does not acdept tabs. Sorry for the eyesore.

Test
1=Zephyr Cape group
2=control group

z= (p1-p2)/[pq((1/n1)+(1/n2))]^.5
p1=27/300
p2=26/300
p=53/600
q=547/600
n1=300
n2=300
z=.14386

This is less than 1 standard deviation from 0, where the hypothesis was that p1-p2=0. They are essentially the same (Zephyr Cape vs White Cape).
You can see from the chart above that when the Zephyr Cape was worn there were 27 misses. When the White Cape was worn, there were 26 misses. So 1 more miss occured during the 300 attacks. Is this statistically significant? No. It failed the z test. Statistically speaking, the Zephyr Cape and the White Cape provide the same protection of MBlock. *Note however the White Cape's +5 DEF does provide 5 points of defensive benefits once hit in additional to MBlock protection.

So in conclusion, the Zephyr Cape provides no additional protection other than its +10% MBlock. So when the extra 10% MBlock (from the Zephyr Cape) causes a miss, then the RedCape graphic appears. It is nothing more than an extra graphic for a normal MBlock miss.

This post has been edited by Detah on 14th December 2004 17:21

I am officially touched that my query has made such a response. thank you very much!
one nitpick though: locke doesn't have the lowest mblock. edgar, cyan and setzer all have 1%. at any rate though, it has been very helpful. thanks again. personally I would have been grateful for my name to have come up but i don't care

--------------------

"Fire and steel
follow me through the lands
you will burn hordes of hell
in the deadly raging flames of revenge"

Rhapsody - Flames of Revenge

Someone has learned statistical maths! Congrats!

Your test, though, is somewhat flawed. FF3e has a well-known bug where evade is never taken into consideration. While your test is valid for the actual in-game effect, it fails to take into account the intended purpose of the cape.

Not that it makes it any less right, mind you.

--------------------

"Judge not a man by his thoughts and words, but by
the quality and quantity of liquor in his possession
and the likelyhood of him sharing."

Yes. Locke does not have the lowest MBlock. I remember now why I chose him.

I created a new game just for the purpose of testing the redcape feature. So my charge was to acquire a Zephyr Cape and a White Cape with the smallest amount of changes to the characters. I wanted the smallest changes to the characters to get the cleanest Control Group that I could create. So I did not increase any of his stats in any way, other than giving him the equipment you see. I gave him defensive equipment to increase the amount of time I could stay in battle without dying. Speed was also an important factor, because I found myself healing just about every round. Locke has the highest initial Speed score, plus the Sneak Ring gives him a nice Speed bonus as well. Keeping in mind the goal of a clean Control Group, I completed my test as early in the game as possible. That is right after the Magitek Research Facility scene and you acquire the airship. Sufficient to get all the eq and the airship to get back to the Leafer / Dark Wind area in front of Narshe. I should have had access to Cyan, Edgar and Setzer at that point, but they were not the best test subjects because of their poorer Speed stats and armor selections. A low MBlock was helpful to my design, but not necessary. So I settled for 2%. But I appreciate the correction. One of these days I may even complete my Character list as well.

The longer I live, the more I am astounded by the lack of proper math skills by the general populace. This is basic probability and statistics, folks. You should have learned this by the time you leave high school. If you didn't, you should write a letter to your local school board and complain about the substandard education that you received.

I never claimed to be testing the Evade trait. The original purpose of the experiment was to determine if a) the redcape graphic occurred more than the +10 MBlock alluded to, and b) whether the Zephyr Cape blocked more than its stated 10%^ MBlock, as its descriptions insinuates. Someone on another message board, stated that the redcape graphic appeared to block 30% of the total attacks (not 30% MBlock, mind you) directed at them. I thought this was extraordinarily high and unbelievable, so I created my experiment. The experiment is designed to test only the +10% MBlock trait of the Zephyr Cape and the frequency of the redcape graphic; nothing more. If you believe my experrimental design is flawed, I would appreciate something a little more detailed than to note that I didnt account for something that was not part of the experiment. I did not account for sunspots either; but I should have stated that earlier.

My test is not flawed. It is a very clean experiment. It is well documented that the Evade trait/function does not work. So it was not necessary for me to point this out in my test results, nor to account for it during my experiment. Evade simply has no bearing on my design or the results.

This post has been edited by Detah on 15th December 2004 06:18

Evade may not have any bearings on your results, but the fact remains that the intended use of some items relies on a stat that was removed from the game.

You may want to read my post again; I was not implying that the results of your experiment were flawed, but that the theory behind it was due to a bug in the game.

--------------------

"Judge not a man by his thoughts and words, but by
the quality and quantity of liquor in his possession
and the likelyhood of him sharing."

Siverlance:
The evade bug doesn't really have anything to do with this test.

Detah:
I'm not sure your test results show what you are implying. From what it looks like, you've proven that the zephyr cape and white cape function similarly in terms of miss %, a fact we already knew. What you were claiming to test was the % of time the red cape graphic appears and to do that you would need to compare a character w/ Mblock +10 vs. a character without. Or alternatively the % of times you see the graphic vs. not seeing the grapic on a miss.

This post has been edited by The Ancient on 15th December 2004 20:03

--------------------

"That Light has bestowed upon me the greatest black magic!"

Let's see.

The goal was not to simply measure the frequency of misses from 10% MBlock. The goal was to determine 2 things. My hypothesis was that a) the redcape graphic is simply the Zephyr Cape's representation for 10% MBlock. Cool lookin' but its just like any other piece of equipment with 10% MBlock. And b) does the Zephyr Cape block more attacks than what is shown in the stats window (ie +10% MBlock)?

So lets see.

I took the number of misses with the White Cape, which we know only has +10% MBlock and compared it to the total misses (redcape or normal) when wearing the Zephyr Cape. 300 attacks under each scenario. 26 misses with White Cape and 27 misses with the Zephyr Cape. Then the appropriate test for significance is the z test. It failed the z test. So I concluded that the Zephyr Cape has statistically the same 'blocking ability' as the White Cape. I'm 99.99% sure that my test does exactly what I stated.

So I did what I wanted. It wasnt meant to do anything more than that.

Now one could also set up this test to test the 'blocking ability' of any old +10% MBlock item. I dont think there is any value in this, because several people here have broken open the rom cart and looked at the code for MBlock. 10% MBlock will block 10/128=7.8% of attacks.

One question remains.

Why go through all that trouble? Disassemble, peek at source listing, get answer. End of problem. Probably in the C2 bank, or maybe C1. C2 contains the code for most actual effects, but C1 handles a lot of graphical stuff in battle (as well as battle event scripts and animation scripts). Should be a piece of cake.

You've got one heck of a lot of patience, though, to go through all that testing.

--------------------

"Judge not a man by his thoughts and words, but by
the quality and quantity of liquor in his possession
and the likelyhood of him sharing."

Quote (Silverlance @ 15th December 2004 15:50)

Why go through all that trouble? Disassemble, peek at source listing, get answer. End of problem.

Sure, but if he already knows statistical math and not 65816, statistics may be a faster route.

Of course, the fastest way would have been to just ask.

--------------------

The Unoriginal White Sheet

-Master ZED

Quote (Master ZED @ 15th December 2004 21:53)

Of course, the fastest way would have been to just ask.

Meh, why spoil his fun, Zed?

And anyone who doesn't know 65816 assembly should be ashamed. Of course, it goes without saying that my view of the world is extremely wonky and should never be trusted or taken seriously under ANY circumstances. Unless the pink polar-toffee bears DO attack. Then I'll gloat.

Wait. Forgot. This isn't a ROM hacking board. Ignore the above statement. Although my beliefs hold true.

--------------------

"Judge not a man by his thoughts and words, but by
the quality and quantity of liquor in his possession
and the likelyhood of him sharing."

Well, I used to know 80x86 assembly, does that count?