Tripping, Not so random?

[SCOTU]

While listening to Youko/SamuraiPanda's SMYN webcast thingy the other day, a thought occurred to me based on the topic they were covering. It's obvious that replays simply store the inputs given to the characters each frame. The game then runs this input data through its game engine to generate the replay. Well, it turns out, if you have a hacked wii running the NO TRIPPING code and save a replay, you can watch your replay and there's no tripping... so long as the wii that you're watching the replay on has the NO TRIPPING code on. If you're watching a game of NO TRIPPING code on a wii that doesn't have the NO TRIPPING code on, then at some points the replay will differentiate from how the match was actually played. This occurs when someone "randomly" trips in the replay (but wouldn't've in the game they played). This also made me realize that the game engine must have some DEFINITE WAY of calculating when to TRIP, because non-hacked replays are always consistent with the matches that the replays were based off of. This means that the tripping is NOT RANDOM with at least respect to something. Because of this, tripping may have a "chance" of activating each action that could cause a trip, but if the RNGs (random number generations) are the same every game, it may be calculable to KNOW when someone's going to TRIP.

I will now add the disclaimer that I did in fact think this through decently well, and realize that there is probably a different random seed for each game, and that that seed is stored in the replay file to recreate the match exactly. However, this is not necessarily the case. There could be only one random seed (unlikely), or there could be a simple pattern to the random seeds, and we could have a "Press Your Luck" style random.

 

[MuBa]

According to the "modifier" we have tripping is actually a random action made by the coding in game. Much like one of those electronic coin flippers/dice rolls where numbers will randomly generate.

Tripping (Non-attack induced) always start at the initial dash animation. So basically when the trip generator turns on, the initial dash animation is temporarily replaced with the tripping animation.

I believe tripping is as random as Luigi's misfire.
.
.
.
.
But on the other hand there is a weird thing that happened with the CPU. Just for the heck of it I tried using a Regular Brawl Replay on a Hacked Wii, and obviously expecting a different match outcome because of the lack of tripping, I've noticed that when my character dies the CPU taunts, which never existed in that particular time. Think you can explain this unusual phenomenon?

 

[SCOTU]

as i was saying in my post, i Know that tripping is generated by an rng, and there is a "chance" of it. What i was hypothesizing is that there may be only one random seed, or only a few.

Perhaps replays don' store cpu inputs, but rather just run the simulation through to the cpu AI and get new inputs, since the cpu's should do everything the same given the same environment.

 

[Pierce7d]

This doesn't make sense either though, because last I checked, CPU's use mix-ups. Otherwise, D3 would always perform the Exact same follow up out of a chainthrow. Start a match with a CPU, and don't press anything. Each match will not be exactly the same. CPU's pick an action from a choice of possible actions that make sense, with a set of limits. I did previously consider what you suggested Scotu, but I think the replay just saves certain random numbers, like tripping, and player start positions (because each slot doesn't start in the same place), etc.

 

[leafgreen386]

In all likelihood, every time a button is input it influences the RNG, and it's also likely that the number of the frame the button is pressed on also has an influence (note that "button" here can refer to the control stick as well). Not inputting any button inputs for a match doesn't create identical scenarios because the game has plenty of other sources of entropy between matches, and even during the match for what the CPU is doing.

Just based off of what I know of some other games' RNG systems.

 

[Levitas]

This doesn't make sense either though, because last I checked, CPU's use mix-ups. Otherwise, D3 would always perform the Exact same follow up out of a chainthrow. Start a match with a CPU, and don't press anything. Each match will not be exactly the same. CPU's pick an action from a choice of possible actions that make sense, with a set of limits. I did previously consider what you suggested Scotu, but I think the replay just saves certain random numbers, like tripping, and player start positions (because each slot doesn't start in the same place), etc.

Not necessarily. If you seed the beginning of the match identically, the RNG value should be the same if the CPU checks at the same times, causing it to perform the same actions it did in the game.

 

[SCOTU]

also, since replays seem to just run the situations through the cpu ai instead of saving all the cpu inputs, (as proven by the taunts that didn't originally take place). I believe that any "mixups" that cpu's do are merely situational differences (%, position, character, etc) or are based off the random seed.