Ice
In short:
- always same facing (ignores character facing)
- half x-velocity of itr
- fixed amount of ice particles
- fixed frame start ids
- start velocities (x,y) is random
- start position after first TU is random
- normal gravitation

I talk about the particles/ice chunks while breakage of the frozen sprites (201-202).
The first facts: it is NOT randomness of: amount of particles and action id. That means, the state 15 triggers the spawn of exact 15 "broken ices" (my calling for now). Each time it spawns:
2x the frame 120
3x the frame 130
4x the frame 125
6x the frame 135
total: 15 (yeahhh i can count)

Everytime the broken ice spawns to the position X: 0, Y: 0, Z: 0 - !relative! to character (here: frame 202). The wait-value equals the original wait-value in "broken_weapon.dat" (no randomness). Now a little bit special: after the first TU, the objects get a random position - independent of the random velocity.
Also important: the broken ices don't inherit the DvX/Y of the character frame. In other words, DvX/Y in the frame 201 and 202 has no effect to the broken ices. More interesting, they get the half x-velocity of an itr. Y will ignore. That means attacks like Tiger Dash of Woody have an effect to broken ices.  The facing of a broken ice is always "normal" (like a character with the view to the right side). No alignment to the frozen character.
The default range without influence from other object is between -5 to 5 for x-velocity and -8 to -17 for y-velocity. The start position is between -19 to 19, the velocities were deducted already!!!
To understand my charts: the charts have on the left side the probability of incidence the velocity/position. On bottom is the pixel value. The bar chart is in percent but over a bar you can see the amount of incidence of this value.





Here you can see, that the default ranges without influence from another object are: -19 to 19 for x-start, -5 to 5 for x-velocity, -28 to 0 for y-start and -17 to -8 for y-velocity.

In the next charts do you can see the influence of Tiger Dash:

Firstly, LF2 distinguishes between "in-state-18" and "after-state-18".

- in state 18 -
In short:
- always same facing (ignores character facing)
- 25% chance to spawn in each TU
- inherits x-velocity of character
- start x-velocity value is random
- actually y-velocity value is -1 (each time)
- start position after first TU is random
- normal gravitation

"Burning Smoke" particles have similarities to "broken ice" parts. They spawn to the position X: 0, Y: 0, Z: 0 - !relative! to character -, get a random position after the first TU - independent of velocities. The wait-value equals the original wait-value in "broken_weapon.dat" (no randomness).
BUT the spawning behaviour looks very strange in the first time. To be honest, it is really easy, if you check the singles TUs. To 25% probability spawns a particle per TU. Each time the particles start by the frame ID 140.
The range of the random x-start-position is between -29 to 29. Y-start-position between -29 to -1. Random of x-velocity is between -5 to 5 like "broken ice". In comparsion to "broken ice", the x-velocity of the character has 100% influence to the particle.
Y-velocity starts always by -1 and gets the dvY-value from frame (default: -1, -2, -3, -3, -2).








By hitted from a fire ball:


Perhaps do you ask you now, if the last chart wrong or in contradiction to my results. Answer is nope. The last chart shows the influence of fire balls. While the hitting of the fire ball, the character is in hit-lag/shaking. So the character doesn't have any velocities. Burning Smoke particles can spawn already while the shaking. That is the reason for the low bars of -5 to 3 (-5 to 5 is the normal range). After the shaking frames, the character gets velocity and also the particles. The amount of TUs while shaking are very low in comparsion to total 16 "fire-flying-frames" (I love my 'neologism'). That is the reason for the first low bars.


- after state 18 -
In short:
- 5-7 particles (dependent of previous state-18-frames)
- else like "in-state-18"

After the character was in state 18 (it is NOT recommend that its goes into falling state/frame), it spawns between 5-7 particles. This behaviour its only in the first none-state-18-TU. That means it is not possible that it spawns one more particle, after the first TU. The amount of spawning of particles are different and dependent of the amount of previous state-18-frames... Between 1-8 TUs, 7 particles spawns to 100% (not variable).
Randomness of start position, velocity, inherits of velocity, like in-state-18.

The probability is too stupid for me if 5,6 or 7 particles. But I want give a benchmark: 7%, 34%, 59%.

These particles are a little bit easier.
In short:
- always same facing (ignores character facing)
- fixed amount of stone particles
- spawns of an object between 0-3 (5 times)
- spawns of an object between 4-7 (8 times)
- start position is random
- start velocities (x,y) is random
- ignores character velocity (x and y tested)

While a stone breaks down, there are spawns a fixed amount of particles. It spawns 5 particles which frame ID is between 0-3 and 8 particles between 4-7. In contrast to "Broken Ice", the particles don't spawn on x: 0, y: 0 (relative to stone). They have already a randomly start position. The range of the x-position AND y-position is between -3 and 3. By spawning, the x-velocity is between -5 and 5 and y-velocity -17 and -8. In case of a one-hit attack with dvX: 20 and dvY: -20: no influence on the start position and velocities.
Same like all other: ignores the "hit facing" (the trigger).

In short:
- like stone but
- fixed amount of particles are 15
- spawns of an object between 40-47 (5 times)
- spawns of an object between 50-57 (10 times)

Equals "Stone" but fixed amount of particles are 15 and the subdivision of the ids are different. Else read "Stone".

In short:
- always same facing (ignores character facing)
- fixed amount of particles
- start position is random
- start velocities (x,y) is random
- ignores character velocity (x and y tested)

While a stick breaks down, spawns 5 particles. The frame ID of these particles are between 10 and 17. Similar to "Stone" the particles spawn not on the coordinates 0,0,0 (relative to object). They have already a randomly start position. The range of the x-position AND y-position is between -3 and 3 (like stone). But the velocity range of x is between -5 and 5 and y-velocity between -9 and -6.

In short:
- like stick but
- 7 fixed frames
- spawns of an object between 20-27 (5 times) //stick of the hoe
- spawns of an object between 30-33 (2 times) //blade of the hoe

Equals "Stick" but fixed amount of particles are 7 and the subdivision of the ids are different. Else read "Stick".

In short:
- like stick but
- has no own particles
- shares the particles of "Wooden Box" and "Hoe"

Equals "Stick" but it spawns two objects with the frame ID between 54-57 and one object with 30-33. Else read "Stick".

In short:
- like stick but
- 4 fixed frames (between 60-63)

Equals "Stick" but it spawns one more particle. Else read "Stick".

In short:
- like stick
- frames between 170-173
Equals "Stick". Else read "Stick".

In short:
- like stick but
- 9 fixed frames
- 1x 70-73
- 2x 80-83
- 6x 74-77

Equals "Stick" except for y-velocity. Else read "Stick".
The y-velocity is "default" (like "Stick" etc.) for 70-73 and 80-83. Though 74-77 have a special y-velocity. Their range is between -12 and -4.

In short:
- like stick milk
- 9 fixed frames
- 3x 160-167
- 6x 74-77

Equals "Milk" except the subdivision of the IDs. Else read "Milk".
The frames 74-77 have the special y-velocity (like milk). Their range is between -12 and -4.