Talk: Wonder Square (Final Fantasy VII)

The image for the 4th 3D battler is wrong.

Wonder Catcher Minigame

I was playing the aforementioned minigame (left machine) and I got a prize of 80 GP, which is not listed as a prize in the article on this game. Is that a mistake in the information? Arcus II (talk) 22:56, March 1, 2015 (UTC)Arcus_II

Mistake in the article, it said "100 GP" which is not possible from the left machine. It is indeed 80 GP. Corrected. JBed (talk) 23:05, March 1, 2015 (UTC)

Wonder Catcher Values

I am wondering where did these values come from? Did someone reverse-compile the source and find it out or were they just done with a lot of trials? The left machine adds up to 99.99% (which is within rounding) but the right machine adds up to 100.12% which is way worse than rounding so there seems to be a mistake.

I wrote a script to automate the PC 2012 version of the game that continually plays it and records the outcome. I am currently working on the left machine and up to 3068 (but 8 hours while sleeping should greatly increase it) and my numbers aren't that close to the given ones. If the previous ones are just lots of trials, I may have the number of trials beat soon (all the guides I have seen online say people did hundreds of trials, I will probably soon be at tens of thousands of trials) and would be willing to contribute my numbers (and repeat it for the right machine, although I won't be gathering megalixirs while doing it...).

Thaimin (talk) 08:30, April 23, 2016 (UTC)

That would be excellent. :) I thought the numbers on the wiki were from the game somehow since it mentions the error whee it says you get Ether you actually get Phoenix Down and vice versa, but if they are not representative of what items you actually seem to get, getting the tested numbers would be good. We should mention though the method they were got somewhere, maybe in a footnote.Keltainentoukokuu (talk) 16:24, April 23, 2016 (UTC)

They are from the data. If you look at the edit form you will see the actual values.

The game randoms a byte (a value between 0 and 255), and then does if-less-than and else to determine the outcome.

Left machine:

• Potion (100/256)
• Nothing (75/256)
• 1 GP (50/256)
• 3 GP (23/256)
• Phoenix Down (receipt: Ether) (5/256)
• Megalixir (2/256)
• 80 GP (1/256)

Right machine:

• Potion (100/256)
• Nothing (75/256)
• 1 GP (50/256)
• 3 GP (20/256)
• Ether (receipt: Phoenix Down) (8/256)
• Elixir (2/256)
• 100 GP (1/256)

It's possible that our %-calculation function is broken somehow, but I can't be bothered to check. All the results seem roughly right at a glance anyway. Oops, one value was rounded to 0dp. Fixed.

It might be better if we list the fraction out of /256 on the page. JBed (talk) 17:41, April 23, 2016 (UTC)

Looks much better now! Thanks!

Just to write down the information somewhere, at 10,004 runs of the left machine I got:

• Potion 3881 38.8% [expected ~3908, -0.7%]
• Nothing 3033 30.3% [expected ~2931, +3.5%]
• 1 GP 1830 18.3% [expected ~1954, -6.3%]
• 3 GP 948 9.48% [expected ~899, +5.5%]
• Ether 212 2.12% [expected ~195, +8.7%] (actually Phoenix Down)
• Megalixir 64 0.64% [expected ~78, -18%]
• 80 GP 36 0.36% [expected ~39, -7.7%]

I was apparently very unlucky, getting far fewer Megalixirs and about 220 GP less than I expected. The potions were basically dead on while the Phoenix Downs and nothings were much more.

How far these are off after 10000 trials makes me suspicious of the quality of the RNG in the game. It I could remember more of my statistics I could report the probability that I would be so far off, maybe another day... 07:27, April 24, 2016 (UTC)