Has anyone gotten their Chocobo color to Midnight Blue?

Quote:

---

Originally Posted by Donjo

(1) You seem to be postulating that feeding a "Deepen color" fruit to a chocobo does not lower the values of the other two colors at similar rates. ...

(2) When I started on the path to Midnight blue ...
...
This time, Red and Green have risen and Blue has fallen. It was at this point that I realized two things.

1. Deepening fruits increased the target color.
2. Deepening fruits lowered the other two colors.

The act of feeding just Deepen Blue made blue rise while the other two fell. The act of alternating between Red and Green made just Blue fall. With this new knowledge, I whipped up a simple little Java program that simulated chocobo feedings. I did a little research to see if other people were thinking the same things I was about the fruit, and got a couple of threads in these forums that suggested exactly what I thought. Furthermore, they suggested a range of change for the fruits: (2A) 3-6 points in the RGB scale at a time. So, I made a Deepen fruit increase the target color by 3-6(out of 255) and lower the other two fall by 3-6, while Lightening fruit dropped the target color and rose the other two. I used this to chart a series of feedings that would theoretically get my Chocobo close to the RGB values for Midnight Blue: 020, 021, 039.

My well planned series of feedings didn't land me on Midnight Blue, however. They landed me on Void Blue: 014, 035, 061. Note just how close the RGB values are between these two blues:

020, 021, 039
014, 035, 061

From here, I Lightened Blue twice, Deepened Red once, Deepened Green once, and then alternated between Deepening Red and Deepening Blue once. According to the theories you're claiming to be false, this did the following things in this order:

1. Lowered Blue, increased Red and Green.
2. Lowered Blue twice, kept Red and Green relatively constant.
3. Lowered Green, kept Red and Blue relatively constant.
...
(3) This is due to two reasons: the observed slight RNG in fruit effects and the comparatively small difference between the colors. If the color you're aiming for requires a large change in RGB values, it's really easy to land on the wrong one because you cannot be sure of exactly where you are. However, using the more precise method I've described in my previous post will, if it fails on the first round of feedings, probably land you on a color that is very similar in values to the one you want. From there, you just feed it (4) 5 or so more fruits to get to your target. Some of these colors are so close together that this kind of manipulation is necessary to obtain something exact. Having fewer fruits to feed lets you be more certain of your current values.

There are, of course, unanswered questions. If your values are nowhere near an established dye color, how does the game figure out what color you'll land on? Does it just take the closest color? Does it assign it based on the ratio between colors? We'll need a lot of testing to figure that one out. Furthermore, if you happen to be a particular color, does the game consider your RGB values to be the exact ones for that color or is it keeping the values your feedings put it at? I'm currently leaning toward the possibility of the game adjusting the values to whatever color you landed on once the color changes.
...
(5) In conclusion, the "precise" method I am suggesting is a good idea because the best position to be in regarding to obtaining a specific color is to be so close to it that very few fruits must be feed to the Chocobo to push it over. This beats the RNG factor and allows you to obtain a color in problem areas where many colors are very close to each other.

---

(1) What I said is, deepen red does not behave the same as deepen blue or deepen green. I did not say deepen red does not make GB to behave in the same matter in the over all RGB. Feeding 5 apples may only allow to hit 1 feather proc, and very likely not even enough, while feeding 5 berries will often result in 2 feather procs, and maybe even 3.

2. So you are one data point. Can we reproduce exactly what you did? If we cannot, then you cannot suggest you have an exact method. Further reading in your details suggests that your path to your result most likely cannot be duplicated.

2(a) - values between 3-6: Then how do we explain these observed results obtain via exactly one feather proc?
Xelphatol

Initial Color Red Δ Green Δ Blue Δ
Desert Yellow (D) 0 0 0
Millioncorn Yellow -8 24 32
Pumpkin Orange 32 40 16
Sunset Orange 16 24 -16
Blood Red 40 32 16

O'Ghomoro

Initial Color Red Δ Green Δ Blue Δ
Desert Yellow (D) 0 0 0
Ul Brown 40 16 -32
Gobbiebag Brown -8 0 -16
Shale Brown 40 32 24
Goobbue Grey 16 0 -24
Lavender Blue -8 0 -40
Raptor Blue 48 0 -24
Othard Blue 40 40 56

They seems to shot any theories using precise numbers to the moon, don't they?

(2) BTW, I am not disagreeing with your behavior observations here, as I have always said back in the experiment thread, however SE programs their color scale RNG, they still must follow the natural color behaviors. What you observed is exactly that.

(2a) What I do like to point out while we are on this is that, ATM, no one has presented any solid reproducible logic that we can use and duplicate the results from one person to the next.

(3/4) This means any theories/spreadsheet/calculator suggesting/applying exact numbers is misleading and should not be used. Just exactly how you came up with 5 fruits, when solid data tracking has shown that one can obtain a new proc message with 1 fruit at 1 color scale, 14 the next scale, and then 3 on the 3rd scale?

(5) Your precise method is based on RNG math, and will require adjustment as you get closer to the desired color. My suggestion is follow the existing proven PNG path, and adjust the colors according. Both of our methods applies the same approach, get close to your desire color, then play with the RNG, as there is no "precise" method to this madness.

I hate to think this way, but the PNG path then adjust accordingly may be the best method we have ATM until someone comes up with some solid results.

Your attempts to debunk my statements in 2(a) merely illustrate one of the unanswered questions I highlighted up above: I've shown that if your RGB values are very close to a color, you can predict with relative certainty that your result will be that color. This still does carry the caveat that if you're trying to be exact on value movements of 100 or more you'll probably be slightly off because of the RNG. Therefore, the minimum number of feedings to get a desired color is 2, with a hopeful maximum of 3. However, I'm not the only data point here. I've guided two FC members to Blood Red and Dalamud Red respectively with exactly 2 feedings by using the same logic I used to obtain Midnight Blue. I'll agree that once may be a coincidence, but three times means that I could be onto something if I keep testing. I can safely assume that my results are reliably "solid" until something happens that completely proves it wrong.

However, we don't yet know an exact method of figuring out what color you'll be if your numbers aren't very close to any color at all. All you did here was show this.

The biggest difference between my alleged "precise" method and your "follow the PNG" method is time, I think. Both of them can certainly get you to a desired color... eventually. My method, however, is currently 3 for 3 on getting to a desired color in exactly 2 feedings. Accepting the possibility of a 3rd feeding being required, that means it will take a casual feeder about 3 days to get their color. Your method carries with it enough uncertainty that it might be necessary to stop at literally every feather change to make sure you're going in the right direction. Why? Because there are a few too many instances in that PNG chart where single feeds of the same color can lead to one of two different colors(as shown by the same color arrow coming out of the same box twice), and this could completely derail the entire process if you don't catch it right away. This means that it's also possible that you'll end up somewhere that isn't even listed on the chart because any direction could have alternate paths. I can't say that my method is an exact science, since RNG prevents that, but following that color chart is far less exact. Furthermore, the chart doesn't have every color listed, which is currently a detriment. Every feeding is a significant chunk of time, and every mistake compounds it. Even if I need silly amounts of fruit to get to a color my way, I still get there quicker. While it may be possible that we can cheat the system to ultimately require fewer fruit to achieve a target, we'll need a handle on how the game assigns colors when values aren't close in order to do it properly. Otherwise, such roundabout methods of reaching a color could actually require far more fruit than necessary.