thunder rune piece

[KFCrispy]

The item drop rates in the bestiary are incorrect. I'm assuming you took the "drop rate", divided by 255, and then by 3?

It should be drop rate / 100 / 3. The game doesn't do a 255 rand, but a 100 instead. This is based on the PSP version at least, so there's a possibility it may have been different on the PS1.

Not sure what the drop rates should be then. So for Holly Boy, Medicine x6 is currently 1.04%. What is the calculation you say the game is doing?

[KFCrispy]

Here are some corrections for the site:

1. Runes

A. Resurrection Rune
- Yell heals 1/3 max HP, not 33%
- Scream heals 300 HP, not 400HP

B. Water Rune and Flowing Rune
- Water of Kindness: Heals all HP in field, 300 HP in battle.

C. Double-Beat Rune
- No penalty
- Add: Only when there is more than one enemy

D. Wind Rune Piece (Earth Rune Piece on suikosource) and Thunder Rune Piece
- Wind Rune Piece also gives DMG +5% per piece besides DEF +3
- Thunder Rune Piece does nothing

E. Killer Rune, Counter Rune, Hazy Rune
- Hard for me to test these

2. Unite Attack
- Pretty Boy Attack: 3x Damage not 2x

Thanks, these are all updated except for Double-Beat not working if there is only 1 enemy.

Some time ago, someone discussed with me that the character they used with double-beat would attack the same target twice. Perhaps it's based on the character's range, but I am not sure. Eventually I want to test it more.

[Raww Le Klueze]

Not sure what the drop rates should be then. So for Holly Boy, Medicine x6 is currently 1.04%. What is the calculation you say the game is doing?

It's a bit weird. Holly Boy has a 8% chance to drop Medicine, but any drop rate is essentially cut in 1/3 because of how the game does it.

All enemies have 3 item slots that they can drop from, even enemies that only drop one thing will have two empty slots in that case and the game will first roll against which slot to check so there's a 33% chance of any of the three slots being checked.

So for Holly Boy there's a 33% chance it won't even try to drop something, and a 33% chance that it will roll for that 8% and another 33% that it will try to drop a Robe at 6%.

If you divide any drop rate by a third then Medicine would be 2.66% and Robe 2% but that's not taking into account that the game will only recognize one drop from one enemy per battle so even if you had a successful roll for Medicine from one Holly Boy and then successfully rolled for a Robe from another one you'd only get whatever came last.

I don't know where we got the original drop rates from though or how it was calculated or based on, but based on what Julian said it looks like we datamined the right numbers but assumed it was a chance of 8 out of 255 instead of 100 which adds up, if you take Holly Boy's Medicine chance as 8 out of 255 = 3.14% divided by 3 slots = 1.04%

So, based on my math taking whatever % we have x3 then x2.55 should get around the original value. Using Medicine as an example, 1.04 x 3 = 3.12 x 2.55 = 7.956 with the discrepancy being because we rounded off with only 2 digits. Would probably be easier to just look at the data again.

[KFCrispy]

oooooh boy.. these are some great technical finds, but for our basic guides it definitely makes sense to try to come up with a simple drop rate . If someone wants to write up an article, we can publish them as in-depth guides!

[Z3R01337]

If you guys want data for drop rates, I have it available here. https://github.com/ak505188/Suikoden-RN ... enemies.js

Drop rate chance is rate/255. This doesn't take into account slot selection or that it goes through the whole enemy party.

[Raww Le Klueze]

Unless they changed between the PS and PSP port it's rate/100, which is also the case for the HD Remaster, in which case the drop rates will vary between versions.

[like2h]

If you divide any drop rate by a third then Medicine would be 2.66% and Robe 2% but that's not taking into account that the game will only recognize one drop from one enemy per battle so even if you had a successful roll for Medicine from one Holly Boy and then successfully rolled for a Robe from another one you'd only get whatever came last.

https://suikosource.com/games/gs1/guides/bestiary.php

If multiple enemies drop an item, the game picks the first in the list
(as the enemies are listed on the battle screen, NOT by killing order)
and ignores the rest.

Which is correct? Or they just say the same thing?

[Pyriel]

It's essentially the same thing, with some liberties taken to make things more understandable. Listing out chances on this game is kind of dicey, and the numbers on any guides should probably be taken as relative indicators rather than exact values.

The way the game works is that it picks an enemy position as a drop candidate. Then, if memory serves, it rolls chances out of 3 to see which drop should trigger, and then rolls against those chances. I have a post somewhere on the forum that describes it in more detail. Probably in one of the cheat code discussions. I honestly can't remember if it only checks occupied enemy positions, so the chances could ultimately be worse than what I'm describing here. It's a chain of dependent probabilities that varies depending on the makeup of the enemy party, how many drops an enemy has, etc. I know one of the rarer drops (I think in Crystal Canyon or whatever toward the end) is actually much rarer than advertised because you have to get past 2-3 enemies that have no drops at all, then win the 1/3 roll to have something like a 0.5% chance.

[Z3R01337]

Got example code here for how enemy drop is calculated: https://github.com/ak505188/Suikoden-RN ... myGroup.js

Game goes through the enemy party (in order), calls RNG to select slot 1-3. If that slot has a drop, it calls RNG again and gets the remainder of that RNG divided by 100. If that remainder is less than the drop rate of the slot/item, you get that drop. If not, it continues to the next enemy. Repeat until drop or went through all enemies.

[like2h]

Got example code here for how enemy drop is calculated: https://github.com/ak505188/Suikoden-RN ... myGroup.js

Game goes through the enemy party (in order), calls RNG to select slot 1-3. If that slot has a drop, it calls RNG again and gets the remainder of that RNG divided by 100. If that remainder is less than the drop rate of the slot/item, you get that drop. If not, it continues to the next enemy. Repeat until drop or went through all enemies.

Nice! That sounds like a simple and reasonable way to calculate drops.

I imagine it probably works the same way in S2. As for the order, maybe it checks enemy position like:
4 1
5 2
6 3

[Z3R01337]

Nice! That sounds like a simple and reasonable way to calculate drops.

I imagine it probably works the same way in S2. As for the order, maybe it checks enemy position like:
4 1
5 2
6 3

I think it is something like that. Easy way to check though is just go through the enemy list via target selection in a fight. The first enemy is slot 1, 2nd is slot 2, etc.

[Pyriel]

It can work, but like I said, it actually hides the real probabilities from the player and the developer alike. I could swear it doesn't keep going after it hits, but maybe the situation I was thinking of there were 3-5 of one enemy and 1 with a much lower chance of dropping something rare. So you have to get past a few rolls for trash.

The second game is a bit more fair, and easier to assign probabilities to. Based on chances out of 255, all the drops are assigned a range of values. The first max value the roll is less than wins. It goes in order of what you kill, so you can increase your chances just in general, and it also allows the Hunter Rune to work properly. Although, the Hunter overrides the default behavior a bit. For example, if an enemy drops Medicine 5/255 and Mega Medicine 5/255, when it dies, anything under 5 on the roll will drop Medicine, anything at least 5, but under 10 will drop Mega Medicine. Anything 10 or higher drops nothing.

[like2h]

For example, if an enemy drops Medicine 5/255 and Mega Medicine 5/255, when it dies, anything under 5 on the roll will drop Medicine, anything at least 5, but under 10 will drop Mega Medicine. Anything 10 or higher drops nothing.

You mean they set a value of 5 for Medicine and 10 for Mega Medicine, and as it checks Medicine first, then Mega Medicine, so both items end with a possibility of 5/255?

Killing order seems a bit odd. What if I kill all the enemies at the same time? Do drops then go by the list order?

[Pyriel]

Yeah. Each drop's value is its chance + the chances of all prior drops in the list. With 3 drops and chances of 5/255, 3/255, 10/255, the checks would be <5, <8, and <18 respectively.

I don't remember exactly, but I believe if you use a spell or unite or whatever to hit multiple enemies, killing more than one, it just tries to get a drop according to their position in the party. There's not much else that would make sense. As I recall, all of this is done during rounds. When damage kills an enemy, the game checks their drops, if a drop hasn't been successful previously. Something that kills multiple targets would just check drops for enemies in the order the damage is applied to them.

Getting drops on-kill at least gives the player some agency, once they know that's how it works. Using the Suikoden 1 method, you're always at the mercy of how the devs structured the party. Also, as a developer, if you want a mosquito to drop Medicine 50% of the time in S1, all their drops have to be Medicine with a 50% chance, or some other janky setup, and they need to come early in the party formation so they're not losing to another enemy all the time.

[like2h]

OK. Thanks for the detailed explanation. It's interesting and insightful. Now, I agree with you that they found a smarter way to handle multiple drops in S2.