Turn order for new enemy spawns

[srd_27]

On this game, every character will actually move once per turn.
When new enemy spawns, they will only be available to move at the next turn, after every other existing characters have moved on their current turn.

This allows you to manipulate the turn ordering of new enemy spawns. I find it hard to explain in words, so I'm just going to use the graphics below.


It lets you choose, if you want these new enemies to move before us (Scenario 1). Or to move after us (Scenario 2), which might allow us to kill the new enemies before they even get to move.

Anyway, I've found this from my recent project of routing a Suikoden Tactics NG+ speedrun. I've been chatting about the game mechanics on the Suikoden speedrun discord, feel free for any Tactics enthusiast to join.
https://discord.gg/pES93ay

[Wolkendrache]

Global turns don’t exist in this game, I guess you used the term for the purpose of illustration. Otherwise, by the time Mizuki is in turn 6, Eugene might be in turn 4.

Reinforcements are queued at the end of the action bar, and I think they are treated as if they had simply waited on the spot previously (meaning without time lag). Certain actions will cause a time lag. The queuing / action bar is determined by 1. Spd stat and 2. time lag by previous action.

If your characters in scenario 2 „turn 2“ had attacked something or used powerful spells, they would probably come after the new spawns in the next „turn“. Likewise, in scenario 1 „turn2“, if there’s a very fast character like Mizuki (instead of the slow one) right after the spawning, and she’ll just wait on the spot, it’s possible for her to overtake the spawns due to her high Spd stat.

[srd_27]

Global turns don’t exist in this game, I guess you used the term for the purpose of illustration. Otherwise, by the time Mizuki is in turn 6, Eugene might be in turn 4.

Reinforcements are queued at the end of the action bar, and I think they are treated as if they had simply waited on the spot previously (meaning without time lag). Certain actions will cause a time lag. The queuing / action bar is determined by 1. Spd stat and 2. time lag by previous action.

I'll prove my theory above here. The two clips in this vid uses the same save state.
https://www.youtube.com/watch?v=S0eXK0BPYwA

First clip shows Scenario 1, where Andarc & Jeremy got queued after the new soldiers, because a slow level-6 Seneca had her last move on "Turn 2".

Second clip shows Scenario 2, where Andarc & Jeremy got queued before the new soldiers. I tweaked this turn order by having Kyril kill Seneca, causing Simeon's spell to be the last move on "Turn 2".

If your characters in scenario 2 „turn 2“ had attacked something or used powerful spells, they would probably come after the new spawns in the next „turn“. Likewise, in scenario 1 „turn2“, if there’s a very fast character like Mizuki (instead of the slow one) right after the spawning, and she’ll just wait on the spot, it’s possible for her to overtake the spawns due to her high Spd stat.

That's what I've thought at first, before realising global turns does exist in this game. I've noticed similar patterns on plenty of different fights as well, where the turn order for new enemy spawns got changed depending if I put a very slow character or not.

[Wolkendrache]

The possibility of lapping others in the action bar (after some time, not that early but still) disprove the existence of global turns. Anyway...

So what’s happening? Imagine a timeline 0-100...

Scenario 1
1. It’s Simeon’s turn so he’s at 100; Seneca is at 95; there’s still a big gap to the next characters because we’re so early in the battle, so those 4 enemy soldiers are at about 55 and then Andarc at 50.
2. Pre-calculation says Simeon will be queued 95 behind Seneca and 50 behind Andarc and so on. Now he’ll execute his action (blazing wall I think), killing all those soldiers...
3. Seneca jumps to 100; Andarc to 55, Simeon from 0 to 5; Kyril and Jeremy are at about 10-20.
4. Enemies spawn 100 behind SENECA but get a bonus for not making a time lag causing action plus good Spd; they start at 30 -> before Kyril, Jeremy and Simeon.

Scenario 2
1. It’s Simeon’s turn so he’s at 100; Seneca is no more; Andarc at 50
2. Pre-calculation says Simeon will be queued 50 behind Andarc and so on. Now he’ll execute his action...
3. Andarc jumps to 100; Simeon from 0 to 50; Kyril and Jeremy from 10-20 to 60-70.
4. Enemies spawn 100 behind ANDARC but get a bonus for not making a time lag causing action plus good Spd; they start at 30 -> behind Kyril, Jeremy and Simeon.

This is just an idea of how it works, but I don’t know any numbers. I think all action bar values are relative to each other. And it doesn’t matter if friend or foe: If that very slow character (Seneca) was an enemy instead, you’d have to kill it just the same to „manipulate“ the queuing of the spawns.

[srd_27]

The possibility of lapping others in the action bar (after some time, not that early but still) disprove the existence of global turns. Anyway...

I tried to test the turn order between Millay with 1 SPD (using Gate of Graska) and a high-level Corselia with 47 SPD.
https://www.youtube.com/watch?v=9WmI9ygMINA

Every couple of turns, Corselia did get double turns. I suppose that supports your theory.

I do feel there's actually a limit on how a high-speed character can outspeed/overlap a low-speed character, which is why you don't see Corselia moving 47 turns before Millay get to move. And also why for most cases, it always feels like everyone always get to move per turn if you don't have any extremely slow characters in the party.

So what’s happening? Imagine a timeline 0-100...

Scenario 1
1. It’s Simeon’s turn so he’s at 100; Seneca is at 95; there’s still a big gap to the next characters because we’re so early in the battle, so those 4 enemy soldiers are at about 55 and then Andarc at 50.
2. Pre-calculation says Simeon will be queued 95 behind Seneca and 50 behind Andarc and so on. Now he’ll execute his action (blazing wall I think), killing all those soldiers...
3. Seneca jumps to 100; Andarc to 55, Simeon from 0 to 5; Kyril and Jeremy are at about 10-20.
4. Enemies spawn 100 behind SENECA but get a bonus for not making a time lag causing action plus good Spd; they start at 30 -> before Kyril, Jeremy and Simeon.

Scenario 2
1. It’s Simeon’s turn so he’s at 100; Seneca is no more; Andarc at 50
2. Pre-calculation says Simeon will be queued 50 behind Andarc and so on. Now he’ll execute his action...
3. Andarc jumps to 100; Simeon from 0 to 50; Kyril and Jeremy from 10-20 to 60-70.
4. Enemies spawn 100 behind ANDARC but get a bonus for not making a time lag causing action plus good Spd; they start at 30 -> behind Kyril, Jeremy and Simeon.

This is just an idea of how it works, but I don’t know any numbers. I think all action bar values are relative to each other. And it doesn’t matter if friend or foe: If that very slow character (Seneca) was an enemy instead, you’d have to kill it just the same to „manipulate“ the queuing of the spawns.

Those numbers are indeed possible explanations on why I can manip the turn order with a low-speed character, without the existence of global turns.

[Wolkendrache]

I do feel there's actually a limit on how a high-speed character can outspeed/overlap a low-speed character, which is why you don't see Corselia moving 47 turns before Millay get to move.

Yes, I've got the same feeling. There's some mechanics going on to prevent too frequent lapping. I haven't seen a fast character moving more than twice in a row before it's the slow one's turn, but I'm not sure (because I always level all characters so their Spd stats don't differ as much as in your example).

[KFCrispy]

srd_27's turn order experiment video is interesting because it suggests what Wolkendrache is saying that there are no global turns, but much of the game actually suggests there ARE global turns.
1. The criteria for a bonus reward for Platinum in the "Give Us A Sign" quest said you need to defeat the Giant Crab "in less than one full turn", though this is a translation from the official Japanese guide. @srd_27, is it possible to load up this quest with your NG+ setup and kill the Giant Crab with the second or last character instead of first?
2. Z3R01337 heavily exploits the "end of turn" situation by casting spells that normally take a ton of time to cast. srd_27 was finding more mechanics on it.

Still, the video is indeed showing that a character can have more actions than another. I don't have an explanation for that. It would be interesting to do more testing, like adding another character to the mix. Millay has 1 SPD, and the global turn theory says she should always get a turn every time Corselia got a turn. What if Millay is actually losing a turn due to the SPD being so low, such as a RNG mechanic that adds or subtracts a little from the character's SPD (SPD=0 no turn?) ?

[Antimatzist]

I also think there's some kind of global turn mechanics at work, but at the moment, they elude me. There's got to be an internal clock of sorts. In srd_27's video, the turn order is VERY consistent with Corselia double and single turns alternating. I tried to put some values for internal clocks there, but it didn't work out. Possibly, it's because the characters got switched in? Also, is Corselia getting a triple turn in the beginning or is it a double turn and you accidentally clicked the "back" button once?

A good battle to check would be Cedric's recruitment battle. A single enemy and you can test out several characters (and take also Cedric's speed into account).

[srd_27]

2. Z3R01337 heavily exploits the "end of turn" situation by casting spells that normally take a ton of time to cast. srd_27 was finding more mechanics on it.

Regarding spell casting, these are the interactions that i've found during my NG+ testing. At first, it made me believe that global turns exist, but Wolkendrache's hypothetical number sequences could also explain these situations without an actual existence of global turns.

On Turn 1, it seems every spell casting can only be finished at the end of turn, no matter how fast the mage is. While on following turns, this doesn't seem to always be the case.
Exception is if we have one character with very low SPD, they might act after the spell casting.


If the mage tries to do two casting spells in a row, he might get the spell casted immediately on Turn 2. But if the mage stands by on Turn 1, he usually gets a delayed cast on Turn 2.
(Note: for a slower-casting spell such as Shining Wind, or for a mage with low level magic skill, the "Turn 2" spell might be casted on early parts of Turn 3 instead)

[srd_27]

1. The criteria for a bonus reward for Platinum in the "Give Us A Sign" quest said you need to defeat the Giant Crab "in less than one full turn", though this is a translation from the official Japanese guide. @srd_27, is it possible to load up this quest with your NG+ setup and kill the Giant Crab with the second or last character instead of first?

So I tried both a 1-move & 2-move fight for Rainbow Shell. We didn't get the Platinum bonus on the 2-move fight, so this requirement is only for a 1-move win.

1-move: https://www.youtube.com/watch?v=0jF6bqhGKjk
2-move: https://www.youtube.com/watch?v=kzTpOkVJPOc

[Wolkendrache]

If the mage tries to do two casting spells in a row, he might get the spell casted immediately on Turn 2. But if the mage stands by on Turn 1, he usually gets a delayed cast on Turn 2.
(Note: for a slower-casting spell such as Shining Wind, or for a mage with low level magic skill, the "Turn 2" spell might be casted on early parts of Turn 3 instead)

Scenario 1: In turn 1 the mage (for some reason the fastest character) is put at the end of the line by casting time. Even with an additional lag for casting a strong spell, the gap to the other characters is still bigger than the casting time in turn 2 (hence casting „immediately“). But keep in mind that the other characters are now short before lapping the mage.

Scenario 2: The situation doesn’t change in turn 1, the mage might probably even get ahead a little by just waiting, but the other characters are still quite close behind. In turn 2, same thing happens as in scenario 1 turn 1. And the mage still has more time until getting lapped.

I think you’d see the downside if you add turn 3 to about 5 with the mage continuing to cast spells. Also, if you want to cast „immediately“ in turn 1, just have the slowest character be a mage.

[srd_27]

An unusual turn order scenario came up in my recent run, vid link below:
https://www.twitch.tv/videos/1928970083?t=01h12m04s

When Corselia is casting Shining Wind, the turn order initially had her turn before one of the enemy archers. But after Jewel/Paula spawns, the turn order get reshuffled and Corselia is now behind all of the enemy archers.


Can't think of a reason for this, don't think enemy's stats got changed after new units are spawned isn't it?

[Wolkendrache]

Hmm...
Ok, we know this behavior is triggered by the appearance of Paula & co. First of all I guess Corselia is the issue, not the 2 enemies getting ahead. The only thing special about her is that she is casting at the time. I don’t think her being a switched-in character has anything to do with it, unless the action bar change had happened immediately after her turn.

So here are some reasons I can think of...
1. The allies‘ appearance cause a slight delay to all of our characters, just enough for those 2 enemies to get ahead of Corselia. It would mean that an intended formula is applied here, but I see no reason for it.
2.1. The allies‘ appearance cause a slight delay to casting characters (mages waiting for the execution of a spell). This would rather be intended, but again, it doesn’t make sense.
2.2. When everyone moves forward in the action bar after the allies appear, casting characters are ignored. This would be unintended.
3. The 2 bowmen are on the same value on the action bar. When Corselia’s casting time delay is applied, it coincidentally places her on the very same value. She’s placed ahead of the bowmen, maybe because she’s an ally or because she’s the acting character (some rule). Now, when the allies arrive, the whole action bar is reprojected without any changes in values, but for all creatures on the same value a rearrangement rule is applied different from the one above...
3a. This time casting characters are reprojected last.
3b. Enemies are reprojected first, then allies.

For 3. to be the case, the 2 bowmen must have identical Spd values in that battle. If you have an in-battle save, can you check that?

[Pyriel]

3c is that they're just sorted in a way that isn't deterministic, or at the very least isn't from the player's perspective. For example, the re-sort starts grabbing from the bottom of the original list, and that unintentionally gives an advantage to characters that were previously lower in the order, if the recognizable/intended criteria are equal.

Edit for a little extra clarity in case it's needed: This is a common problem in sorts. If you give me a randomly ordered list of everything you did today with time down to the hour, I can sort it, but what order do I put the 20 things you did at 1PM in? Without some secondary key, the entries can just come out in seemingly random order based on things like the starting state, and how entries are shuffled in the work space. So the first time the result is "1PM Toilet, 1PM Washed hands" and the next time it's run they're reversed.

[srd_27]

Pyriel's explanation seems to makes sense. Corselia's casting probably happen to have the same "order number" as the two archers. When the new units spawn, the order got reprojected in a way that changes characters with the same "order number" (although who knows how exactly this works).

For 3. to be the case, the 2 bowmen must have identical Spd values in that battle. If you have an in-battle save, can you check that?

Don't have the save for this exact battle. But I have a save state for this battle from another playthrough, with the same setup and strategy. But Andarc & Corselia's SPD value will probably be a bit different.
https://drive.google.com/file/d/15RKcXK ... sp=sharing