Character Growth rates (and Data-mine Topic)

[Sire]

Don't know where I should put this, but I think the Prayer Ring may have a hidden "Miracle" effect.

My Ignatz was going to die due to being doubled, but suddenly he proc'ed Miracle. He doesn't have Miracle as a skill and I never made him into a Priest. However, he did have a Prayer Ring equipped.
- - - - -
For other info, I do have the Crest of Blaiddyd on him and his Battalion are Golden Deer Archers.

[Moonlit Knight]

Don’t know if this has been pointed out anywhere, but the Lost Items are all found in the same exact location where the true owner was standing the previous month (with some notable exceptions: I.e., Bernadetta’s items are just outside her door instead of in her room with her).

Edited August 26, 2019 by Moonlit Knight

[Hauke]

Would it be possible to get the tecture for the Sword of Moralta? Or a better picture of the model? I want to see how it looks to draw it, but ingame it's not that easy.

[Silly]

@VincentASM

Given that you replied to the thread, I'm sure you're aware of this recent reddit discovery about how autoleveled bases work: https://www.reddit.com/r/fireemblem/comments/d0w2hr/an_analysis_on_the_monastery_autolevelling/

I have been testing the above link and for the most part it looks like it works well for all recruitable units. However, I've been seeing a few inconsistencies and given that you've actually looked through the game files I was hoping that you might have some insight into things.

Most inconsistencies I've found are certain non-student units joining with stats that are ~1 point off of from where they should be. For example, using the formula Alois should have 18 base defense when recruited in Chapter 11, but when I recruited him he only had 17 defense. Not sure if there is some unknown stat variation affecting things by about +/- 1 stat point, but for the most part things line up.

I have encountered a few bigger problems though. The biggest inconsistency that I'm seeing is Gilbert. He joins with about 30 defense, but the Fortress Knight class that he's in has a massive 60% enemy defense modifier, which should imply that his base defense on join be about 10 points higher. Flayn (and I believe Hanneman) are also smaller inconsistencies, joining with about 3 point less resistance than the numbers should imply.

Gilbert specifically is a big issue, since the numbers really don't match for him. Not the default enemy fortress knight data, nor the ch 5 fortress knight data that you also have listed. I was wondering if there was some other sort of Fortress Knight growths in the game data that he was using for level ups instead, or he had different bases as an ally unit versus a playable unit, or something else that could explain how his defense is so different from what it should be.

Edited September 8, 2019 by Silly

[VincentASM]

Sorry that I've been kinda quiet. Been busy with various things, while trying to squeeze in as much Three Houses play-time and testing as possible ^^;;

Added the base activation rates to the Crests page:

https://serenesforest.net/three-houses/characters/crests/

@Sire

Thanks! I had a feeling it worked like that. Otherwise it seems weird that the ring is inferior to the Goddess Ring.

@Silly

Hmm, that's curious.

Maybe try plugging in the values for the other Fortress Knight class?

HP: 35 (+5)
STR: 14
MAG: 8
DEX: 12
SPD: 8 (-6)
LCK: 10
DEF: 13 (+2)
RES: 7

Those are the base stats, with the class-change boosts in brackets. I'm assuming Gilbert is in the normal Fortress Knight class when you get him, so ignore the CC boosts. But maybe his stats are calculated using those base stats?

Hmm, but that's only a 4 point difference in Defence, I think?

Anyway, besides that, there shouldn't be any other Fortress Knights. Also, there is another Gilbert, but he has the same bases and growths as normal.

[Silly]

3 hours ago, VincentASM said:

 

@Silly

Hmm, that's curious.

Maybe try plugging in the values for the other Fortress Knight class?

HP: 35 (+5)
 STR: 14
MAG: 8
DEX: 12
SPD: 8 (-6)
LCK: 10
DEF: 13 (+2)
 RES: 7

 Those are the base stats, with the class-change boosts in brackets. I'm assuming Gilbert is in the normal Fortress Knight class when you get him, so ignore the CC boosts. But maybe his stats are calculated using those base stats?

 Hmm, but that's only a 4 point difference in Defence, I think?

Anyway, besides that, there shouldn't be any other Fortress Knights. Also, there is another Gilbert, but he has the same bases and growths as normal.

Here is the "supposed" math for Gilbert.

Personal Base = 5
Personal Growth = 45
Class Growth = 60
Alternate Class Growth = 50 (you have a second fortress knight class listed)
Class Bonus = 10

Implied Level 26 Defense: 5 + 25 * (0.45 + 0.6) + 10 = 41.25
Implied Def (other growths): 5 + 25 * (0.45 + 0.5) + 10 = 38.75

Actual Level 26 Defense: 30

The above formula works out relatively well for most other units (within a 1 point difference of their actual bases in most cases). It even works really well for all of Gilbert's non-defense stats. It just for some reason ends up incredibly off with Gilbert's defense.

Either Gilbert's class defense growth is actually about 10-20%, or something weird is going on.

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

EDIT: Okay so after giving this some more thought I think I've stumbled across the correct answer.

Defense and resistance growth is capped at 60%.

If we assume that to be true, this fixes the issue with Gilbert's defense, as well as Flayn and Hanneman's joining Res, and corrects the minor 1 point difference in Alois + Seteth's base defenses on recruitment.

I assume something similar might be contributing to the other places where the stats are inconsistent, but will have to do further research later.

Edited September 10, 2019 by Silly

[Druplesnubb]

Spoiler

On the recruitment page the part where you can recruit Lysithea in Chapter 14 on Crimson Flower needs to be added. Maybe also the part with having to re-recruit Ashe and Lorenz on some routes as well.

Spoiler

As for the crest page, I haven't gotten there yet but I've heard that you fight Nemesis and the Ten Elites on Golden Deer. I don't know whether they have their crests or not but if they do they should be added to the page.

 

[VincentASM]

@Silly

Huh, that's interesting. I guess for high defence characters, it's better to recruit them earlier? With Gilbert you don't have a choice though!

@Druplesnubb

Oh yeah, I'll add the Part 2 recruitment conditions soon-ish.

You are correct about the second part, although I'm unsure how to add them without it being too spoilery. I think there are a few more enemies with crests, but they aren't particularly memorable (or just plain spoilerific). I might have to write a disclaimer in there or something ^^;;;

[Silly]

@VincentASM

Also I noticed a minor error on the website when doing research for the base stat formula.

On your site, Pegasus Knight is listed as having a +2 Cha modifier in the class stat bonuses section. This should actually be a +2 Res modifier.

Edited September 11, 2019 by Silly

[VincentASM]

22 hours ago, Silly said:

@VincentASM

Also I noticed a minor error on the website when doing research for the base stat formula.

On your site, Pegasus Knight is listed as having a +2 Cha modifier in the class stat bonuses section. This should actually be a +2 Res modifier.

Thanks, should be fixed now! Of course, I put the number in the wrong column ^^;;

[Cysx]

On 9/10/2019 at 9:44 PM, VincentASM said:

Added the base activation rates to the Crests page:

https://serenesforest.net/three-houses/characters/crests/

.

Welp, I just wasted my time then, for the most part. It's really nice to have all those odds, though, I only got the combat ones on my own.
Though there is one thing that doesn't match about the Crest of flames; in 500 attacks, the preventing counters and +5 damage effects activated 12 and 16 times respectively, and never separately from the healing, or together. The healing activated 103 times, which matches... but as a result, the other two effects should be much, much rarer than 1/4.

Also an additional tidbit, the crest of Indech cannot activate off of brave effects at all, from what I've tested.

Outside of that, I have a lot of raw exp data that I'm having trouble extracting a formula from, anything you've seen within the game's data that could help, by any chance?

 

[VincentASM]

On 9/12/2019 at 5:15 PM, Cysx said:

.

Welp, I just wasted my time then, for the most part. It's really nice to have all those odds, though, I only got the combat ones on my own.
Though there is one thing that doesn't match about the Crest of flames; in 500 attacks, the preventing counters and +5 damage effects activated 12 and 16 times respectively, and never separately from the healing, or together. The healing activated 103 times, which matches... but as a result, the other two effects should be much, much rarer than 1/4.

Also an additional tidbit, the crest of Indech cannot activate off of brave effects at all, from what I've tested.

Outside of that, I have a lot of raw exp data that I'm having trouble extracting a formula from, anything you've seen within the game's data that could help, by any chance?

 

Hmm, I was wondering if the negating counter effect was separate; I only counted the additional damage, since that was easier to see.

But the 1 in 4 refers to once the Crest has triggered. 16 in 103 isn't too far from 1 in 4, although it is quite a bit lower. Curiously, if you add 12 and 16 together, it's actually close to 1 in 4. Although that's a coincidence.

Sadly, I don't think I'd be able to find any formulae. I can barely identify plain old tables as it is, since none of the files have names (until they get updated in a patch) or labels (well, around 1% of the files do) ^^;;

EDIT

Also, it's never a waste. It's good to have real data to look at, instead of just pulling numbers that could be anything from the code.

Edited September 13, 2019 by VincentASM

[Cysx]

14 hours ago, VincentASM said:

Hmm, I was wondering if the negating counter effect was separate; I only counted the additional damage, since that was easier to see.

But the 1 in 4 refers to once the Crest has triggered. 16 in 103 isn't too far from 1 in 4, although it is quite a bit lower. Curiously, if you add 12 and 16 together, it's actually close to 1 in 4. Although that's a coincidence.

Sadly, I don't think I'd be able to find any formulae. I can barely identify plain old tables as it is, since none of the files have names (until they get updated in a patch) or labels (well, around 1% of the files do) ^^;;

EDIT

Also, it's never a waste. It's good to have real data to look at, instead of just pulling numbers that could be anything from the code.

Ah, okay, the 1/4 odd makes more sense, then. And unfortunately it's really difficult to gauge those two, I'd need to run a set of 5000 attacks to get potent results, and I'll fully admit I don't have the courage for thatXD. Regardless, those are odds low enough that 100% accuracy probably wouldn't help much, it's obvious enough that they're very rare.

That's unfortunate! It's not like I've learned nothing but... the way the numbers grow, I just can't make complete sense of them. Doesn't help that the game does not provide a reliable baseline to work off of, since the prologue exclusively gives you units with a 20% exp boost through their personal... Then you have different tiers of enemies giving different bonuses and I can't even tell if beginner tier or intermediate enemies are supposed to be the base unmodified value, considering the former are rarer... Normal to Hard does modify exp but not consistently... it's all a bit of a mess. And now we have Maddening.

Yeah, you're right. I just basically came in to post my findings, so that was me letting out a bit of frustration.
Though I really need to look into datamining to some extent next time around. Not saying it's easy at all, but I can't help but feel... limited, without it.

Edited September 14, 2019 by Cysx

[Jayvee94]

I dunno if anyone noticed yet.

According to this guide, when a student (or faculty) ask you a question, the person(s) accompanying the student is also motivated.

[Elephantus]

On 8/26/2019 at 2:25 AM, Sire said:

Don't know where I should put this, but I think the Prayer Ring may have a hidden "Miracle" effect.

My Ignatz was going to die due to being doubled, but suddenly he proc'ed Miracle. He doesn't have Miracle as a skill and I never made him into a Priest. However, he did have a Prayer Ring equipped.
- - - - -
For other info, I do have the Crest of Blaiddyd on him and his Battalion are Golden Deer Archers.

I know this is an old post, but I just wanted to confirm this. Happened to me yesterday and the Prayer Ring is really the only logical explanation I could find.

[Cysx]

18 hours ago, Jayvee94 said:

I dunno if anyone noticed yet.

According to this guide, when a student (or faculty) ask you a question, the person(s) accompanying the student is also motivated.

That's correct, assuming they like the answer. You can tell by how they react; unfortunately the students don't like the same things, and the right answer for the asking student will not always be the one others want. This also might affect prof exp to an extent, it's a pretty confusing design choice.

Edited September 30, 2019 by Cysx

[Silly]

Someone on a Chinese forum worked out the formula for certification exam pass rate. The translation was posted on reddit recently.

Apparently, luck affects your exam pass rate...

Obviously this means that luck is now the best stat.

[VincentASM]

Sorry I've been rather quiet. Been busy, while also trying to do as much research as possible ^^

Finished the Adjutants page.

Made some interesting discoveries.

A few people noticed this, but Adjutant Guard prevents enemy follow-ups from killing units.

Meanwhile, Adjutant Follow-Up actually has a decent activation rate, but it seems to be bugged. It doesn't trigger during the enemy phase and support levels C+, B+ and A+ (even if hidden) count as "no support".

[Cysx]

1 hour ago, VincentASM said:

Sorry I've been rather quiet. Been busy, while also trying to do as much research as possible ^^

Finished the Adjutants page.

Made some interesting discoveries.

A few people noticed this, but Adjutant Guard prevents enemy follow-ups from killing units.

Meanwhile, Adjutant Follow-Up actually has a decent activation rate, but it seems to be bugged. It doesn't trigger during the enemy phase and support levels C+, B+ and A+ (even if hidden) count as "no support".

Uh. So follow up was actually meant to be pretty good then. Well, 40% outside of NG+ isn't that high I guess.
In any case, as it is now you spend way more of the game in intermediate + ranks than in the basic ones, so unless they fix it...

Edited October 4, 2019 by Cysx

[VincentASM]

On 10/5/2019 at 12:14 AM, Cysx said:

Uh. So follow up was actually meant to be pretty good then. Well, 40% outside of NG+ isn't that high I guess.
In any case, as it is now you spend way more of the game in intermediate + ranks than in the basic ones, so unless they fix it...

Oh yeah, I was checking some loose ends and unless I'm mistaken, only Byleth has hidden + support levels (and no A+ for same gender characters).

So Byleth is the one who's most affected by this bug.

Then again, the fact it can't trigger during enemy phase seems really weird too. Adjutant Guard works during both phases, so I'm curious if being player phase only is an oversight as well...

[VincentASM]

I've updated the greenhouse page with information from the guidebook.

There are formulae to calculate the yield for any number and/or type of seed, as well as the chance of obtaining stat-boosters.

Just in case, does anyone here know how to write Javascript code (or any online-friendly code) and would be interested in volunteering to create a gardening calculator?

[Silly]

On 10/8/2019 at 5:16 PM, VincentASM said:

I've updated the greenhouse page with information from the guidebook.

There are formulae to calculate the yield for any number and/or type of seed, as well as the chance of obtaining stat-boosters.

Just in case, does anyone here know how to write Javascript code (or any online-friendly code) and would be interested in volunteering to create a gardening calculator?

Is this formula fully accurate?

Chance = Stat-booster co-efficient + (combined [ Grade – 1 ] of all seeds x 5) + (number of seeds x 6) + (cultivation Tier x 5)

Planting five flower seeds and doing max cultivation would yield a combined grade bonus of 50 ((3-1) * 5 * 5), a number of seeds bonus of 30 (5 * 6) and a cultivation tier bonus of 30 (6 * 5), which gives a total chance of a stat booster that exceeds 100%. However, I have definitely failed to get a stat booster from 5 flower seeds + max cultivation before.

[VincentASM]

14 minutes ago, Silly said:

Is this formula fully accurate?

Chance = Stat-booster co-efficient + (combined [ Grade – 1 ] of all seeds x 5) + (number of seeds x 6) + (cultivation Tier x 5)

Planting five flower seeds and doing max cultivation would yield a combined grade bonus of 50 ((3-1) * 5 * 5), a number of seeds bonus of 30 (5 * 6) and a cultivation tier bonus of 30 (6 * 5), which gives a total chance of a stat booster that exceeds 100%. However, I have definitely failed to get a stat booster from 5 flower seeds + max cultivation before.

Thanks for asking!

Yeah, I made a mistake. The (Grade -1) x 5 part only applies for one seed, I believe. That should cap the rate at 100.

So 20 (max co-efficient) + 20 (for a 5-star seed) + 30 (5 seeds) + 30 (pegasus blessings) = 100.

I have no idea what happens during Blessing of the Land. I don't recall getting 2 stat-boosters ever, so I assume it also caps at 100%. But if I'm wrong, do let me know!

If you're mixing seeds, things are a bit more complicated. I guess you'd need to calculate the rate for every type of seed, then somehow split the rate by the ratio of seeds.

Edited October 10, 2019 by VincentASM

[XeKr]

Sorry I’m not very helpful with respect to javascript but here are some probabilities for greenhouse statboosters from the buyable flower seeds.

Adapted flower seed chart from @Aggro Incarnate
Gardening minibooster yield probabilities for STR/MAG/SPD. Gold refers to cultivation cost.

 

Purple (STR)      0G     300G   500G   1000G  1500G  2000G
1 Flower Seed     26%    31%    36%    41%    46%    56%
2 Flower Seeds    30%    35%    40%    45%    52%    57%
3 Flower Seeds    43%    48%    58%    63%    68%    73%
4 Flower Seeds    44%    54%    59%    64%    69%    74%
5 Flower Seeds    50%    55%    60%    65%    70%    75%

 

Yellow (MAG)      0G     300G   500G   1000G  1500G  2000G
1 Flower Seed     24%    29%    36%    41%    46%    51%
2 Flower Seeds    37%    42%    47%    52%    57%    62%
3 Flower Seeds    36%    41%    46%    51%    56%    63%
4 Flower Seeds    44%    54%    59%    64%    69%    74%
5 Flower Seeds    48%    53%    58%    63%    68%    73%
Note:  
1 Southern Fodlan Seed (0G): 26%
1 Southern Fodlan Seed (300G): 31%
2 Yellow Flower Seeds + 1 Southern Fodlan Seed (1000G): 56.3%
4 Yellow Flower Seeds + 1 Southern Fodlan Seed (2000G): 79%

 

Pale-Blue (SPD)   0G     300G   500G   1000G  1500G  2000G
1 Flower Seed     31%    36%    41%    46%    51%    56%
2 Flower Seeds    37%    42%    47%    52%    57%    62%
3 Flower Seeds    43%    48%    53%    58%    63%    68%
4 Flower Seeds    44%    54%    59%    64%    69%    74%
5 Flower Seeds    50%    55%    65%    70%    75%    80%

We also did some calculations to find optimal seed ratios, assuming for different seed combinations each one has the same chance of being chosen for the (Grade -1) x 5 part. We pretty much computed all the combinations but there’s too many to list so here’s some of the more important numbers we found by screening for certain maxima. Note that it’s mainly mixing in Southern Fodlan Seeds (for mag boosters) for total 3 or 5 seeds that has a relevant impact. The grade for Mixed Herb Seeds (for str boosters) is too low, and Angelica/Nordsalat Seeds are not buyable (though can luck into them via shiny spots). 

 

Below are some other optimal seed ratios when maximizing chances at each professor level (meaning using as many seeds and cultivation as allowed for each level). These were computed by a script so sorry for any typos/errors. 

Spoiler

 

Rocky Burdocks (Str). No Angelica Seeds.
The maximum probability for 1 seeds and cultivation level 1 is 26.0%
0 Mixed Herb Seeds. 1 Purple Flower Seeds.
The maximum probability for 1 seeds and cultivation level 2 is 31.0%
0 Mixed Herb Seeds. 1 Purple Flower Seeds.
The maximum probability for 2 seeds and cultivation level 3 is 40.0%
0 Mixed Herb Seeds. 2 Purple Flower Seeds.
The maximum probability for 3 seeds and cultivation level 4 is 63.0%
0 Mixed Herb Seeds. 3 Purple Flower Seeds.
The maximum probability for 4 seeds and cultivation level 5 is 69.0%
0 Mixed Herb Seeds. 4 Purple Flower Seeds.
The maximum probability for 5 seeds and cultivation level 6 is 75.0%
0 Mixed Herb Seeds. 5 Purple Flower Seeds.

Rocky Burdocks (Str). 
The maximum probability for 1 seeds and cultivation level 1 is 34.0%
0 Mixed Herb Seeds. 1 Angelica Seeds. 0 Purple Flower Seeds.
The maximum probability for 1 seeds and cultivation level 2 is 39.0%
0 Mixed Herb Seeds. 1 Angelica Seeds. 0 Purple Flower Seeds.
The maximum probability for 2 seeds and cultivation level 3 is 52.0%
0 Mixed Herb Seeds. 1 Angelica Seeds. 1 Purple Flower Seeds.
0 Mixed Herb Seeds. 2 Angelica Seeds. 0 Purple Flower Seeds.
The maximum probability for 3 seeds and cultivation level 4 is 69.7%
0 Mixed Herb Seeds. 2 Angelica Seeds. 1 Purple Flower Seeds.
The maximum probability for 4 seeds and cultivation level 5 is 79.0%
0 Mixed Herb Seeds. 4 Angelica Seeds. 0 Purple Flower Seeds.
The maximum probability for 5 seeds and cultivation level 6 is 95.0%
0 Mixed Herb Seeds. 5 Angelica Seeds. 0 Purple Flower Seeds.

Premium Magic Herbs (Mag)
The maximum probability for 1 seeds and cultivation level 1 is 26.0%
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.
The maximum probability for 1 seeds and cultivation level 2 is 31.0%
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.
The maximum probability for 2 seeds and cultivation level 3 is 47.0%
0 Southern Fodlan Seeds. 2 Yellow Flower Seeds.
The maximum probability for 3 seeds and cultivation level 4 is 56.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.
The maximum probability for 4 seeds and cultivation level 5 is 69.0%
0 Southern Fodlan Seeds. 4 Yellow Flower Seeds.
The maximum probability for 5 seeds and cultivation level 6 is 79.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

Speed Carrots (Spd). No Nordsalat Seeds.
The maximum probability for 1 seeds and cultivation level 1 is 31.0%
1 Pale-Blue Flower Seeds.
The maximum probability for 1 seeds and cultivation level 2 is 36.0%
1 Pale-Blue Flower Seeds.
The maximum probability for 2 seeds and cultivation level 3 is 47.0%
2 Pale-Blue Flower Seeds.
The maximum probability for 4 seeds and cultivation level 5 is 69.0%
4 Pale-Blue Flower Seeds.
The maximum probability for 5 seeds and cultivation level 6 is 80.0%
5 Pale-Blue Flower Seeds.

Speed Carrots (Spd)
The maximum probability for 1 seeds and cultivation level 1 is 31.0%
0 Nordsalat Seeds. 1 Pale-Blue Flower Seeds.
1 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.
The maximum probability for 1 seeds and cultivation level 2 is 36.0%
0 Nordsalat Seeds. 1 Pale-Blue Flower Seeds.
1 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.
The maximum probability for 2 seeds and cultivation level 3 is 47.0%
0 Nordsalat Seeds. 2 Pale-Blue Flower Seeds.
2 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.
The maximum probability for 3 seeds and cultivation level 4 is 58.0%
0 Nordsalat Seeds. 3 Pale-Blue Flower Seeds.
The maximum probability for 4 seeds and cultivation level 5 is 79.0%
4 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.
The maximum probability for 5 seeds and cultivation level 6 is 90.0%
5 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.

 

 

Optimal seed ratios for highest cultivation level (2000G, Pegasus Blessings) and different seed numbers

Spoiler

Rocky Burdocks (Str)
_______________________
For cultivation level 6

The maximum probability for 1 seeds and cultivation level 6 is 59.0%
0 Mixed Herb Seeds. 1 Angelica Seeds. 0 Purple Flower Seeds.

The maximum probability for 2 seeds and cultivation level 6 is 67.0%
0 Mixed Herb Seeds. 1 Angelica Seeds. 1 Purple Flower Seeds.
0 Mixed Herb Seeds. 2 Angelica Seeds. 0 Purple Flower Seeds.

The maximum probability for 3 seeds and cultivation level 6 is 79.7%
0 Mixed Herb Seeds. 2 Angelica Seeds. 1 Purple Flower Seeds.

The maximum probability for 4 seeds and cultivation level 6 is 84.0%
0 Mixed Herb Seeds. 4 Angelica Seeds. 0 Purple Flower Seeds.

The maximum probability for 5 seeds and cultivation level 6 is 95.0%
0 Mixed Herb Seeds. 5 Angelica Seeds. 0 Purple Flower Seeds.

 

 

Premium Magic Herbs (Mag)
_______________________
For cultivation level 6

The maximum probability for 1 seeds and cultivation level 6 is 51.0%
0 Southern Fodlan Seeds. 1 Yellow Flower Seeds.
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 2 seeds and cultivation level 6 is 62.0%
0 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 6 is 66.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 4 seeds and cultivation level 6 is 74.0%
0 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 6 is 79.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

 

 

Speed Carrots (Spd)
_______________________
For cultivation level 6

The maximum probability for 1 seeds and cultivation level 6 is 61.0%
1 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.

The maximum probability for 2 seeds and cultivation level 6 is 64.5%
1 Nordsalat Seeds. 1 Pale-Blue Flower Seeds.

The maximum probability for 3 seeds and cultivation level 6 is 69.7%
1 Nordsalat Seeds. 2 Pale-Blue Flower Seeds.

The maximum probability for 4 seeds and cultivation level 6 is 89.0%
4 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.

The maximum probability for 5 seeds and cultivation level 6 is 90.0%
5 Nordsalat Seeds. 0 Pale-Blue Flower Seeds.

 

 

The most notable alternatives to just going all flower seeds are in the earlier flower table, but here’s some others with respect to buyable seeds (so mixed herb and southern fodlan)

Spoiler

Rocky Burdocks (Str)

_______________________
For cultivation level 1

The maximum probability for 4 seeds and cultivation level 1 is 46.5%
1 Mixed Herb Seeds. 3 Purple Flower Seeds.

 

 

Premium Magic Herbs (Mag)
_______________________
For cultivation level 1

The maximum probability for 1 seeds and cultivation level 1 is 26.0%
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 1 is 41.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 1 is 49.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

_______________________
For cultivation level 2

The maximum probability for 1 seeds and cultivation level 2 is 31.0%
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 2 is 46.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 2 is 54.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

_______________________
For cultivation level 3

The maximum probability for 1 seeds and cultivation level 3 is 36.0%
0 Southern Fodlan Seeds. 1 Yellow Flower Seeds.
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 3 is 51.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 3 is 59.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.
_______________________
For cultivation level 4

The maximum probability for 1 seeds and cultivation level 4 is 41.0%
0 Southern Fodlan Seeds. 1 Yellow Flower Seeds.
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 4 is 56.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 4 is 69.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

_______________________
For cultivation level 5

The maximum probability for 1 seeds and cultivation level 5 is 46.0%
0 Southern Fodlan Seeds. 1 Yellow Flower Seeds.
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 5 is 61.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 5 is 74.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

_______________________
For cultivation level 6

The maximum probability for 1 seeds and cultivation level 6 is 51.0%
0 Southern Fodlan Seeds. 1 Yellow Flower Seeds.
1 Southern Fodlan Seeds. 0 Yellow Flower Seeds.

The maximum probability for 3 seeds and cultivation level 6 is 66.3%
1 Southern Fodlan Seeds. 2 Yellow Flower Seeds.

The maximum probability for 5 seeds and cultivation level 6 is 79.0%
1 Southern Fodlan Seeds. 4 Yellow Flower Seeds.

 

None for Speed Carrots since Nordsalat isn't buyable.

Really those first 3 charts are the most useful imo.

[SkyOc1]

Does anyone know if the game checks seed by seed for the stat booster items? You can get above 100% chance for a stat booster, especially with blessing of the land. For example: 4 angelica seeds and 1 western fodlan seed Cultivation 6 of course. The angelica seeds all have 100%/130% chance to give a stat booster while the western fodlan seed has an 80%/110% chance to. How does the game determine which stat booster it gets?

Edited October 13, 2019 by SkyOc1