Final Fantasy VIII

2012-08-25 06:28:39 pm

I did, but that was a while ago and the record is 4 minutes faster now. I haven't gotten around to capturing the latest record yet.

Aspirist13

Edit history:

Aspirist13: 2012-08-29 08:27:17 am

2012-08-27 12:25:29 am

Tomato Rice uploaded the 8:52 record here http://www1.axfc.net/uploader/Ca/so/32425 (password:1464263) if you feel like clicking retry every 2min for quite a while to download it.

It took me like 1hr to start the download due to the "download overload queue" they have, so Ill try to mirror it somewhere for anyone else who wants it.

Edit: Mirror on google drive for easier download https://docs.google.com/open?id=0B2tYpdR242weTzgySHVjSmtmUEk

jpic

2012-08-27 02:56:20 am

thank you for the link!! managed to get the download started after about 20mins

Caracarn

2012-08-27 11:43:04 am

oh cool, I didn't know he actually uploaded the video there, I thought it was just the notes chart. I'll have to see whether he uploads his FF9 run also.

FFgamer86

2012-08-27 11:41:02 pm

SDA Speedruns: 1

I successfully downloaded it, will watch from time to time to see what he does compared to my route.

Poxnor

2012-08-30 04:41:15 pm

Moo! Flap! Hug!

If you're playing Triple Triad with, e.g., the same rule, is the combo cascade of flips triggered before or after any regular flips by the first card played?

Example: let's say that my opponent owns all 8 cards on the board, and I have the last move. To make this simple, assume the card in spot 1 is 1111, the card in spot 2 is 2222, etc., with only spot 5 (the middle) being free.

Now, let's say my last card is 3681 (top, right, bottom, left). That card will do two things: flip the card in spot 2 naturally, and trigger a same-cascade starting with the cards in spot 6 and spot 8. My question is: which of those happens first? Does the same cascade play out then the normal flip, or does the normal flip happen then the cascade start? It actually affters the outcome of the play.

If the flip happens first, the card in spot 2 becomes mine. Then, the cascade starts in spots 6 and 8, earning me spot 7 and 3 as well. The cascade halts after that point, though, because there's nothing more to flip.

If the cascade happens first, then I get 6 and 8, plus the cascade through 7, 3, 2, and 1. After the cascade, there's no flip on 2, since it was flipped in the cascade.

If you're playing Triple Triad with, e.g., the same rule, is the combo cascade of flips triggered before or after any regular flips by the first card played?

Example: let's say that my opponent owns all 8 cards on the board, and I have the last move.  To make this simple, assume the card in spot 1 is 1111, the card in spot 2 is 2222, etc., with only spot 5 (the middle) being free.

Now, let's say my last card is 3681 (top, right, bottom, left).  That card will do two things: flip the card in spot 2 naturally, and trigger a same-cascade starting with the cards in spot 6 and spot 8.  My question is: which of those happens first?  Does the same cascade play out then the normal flip, or does the normal flip happen then the cascade start?  It actually affters the outcome of the play.

If the flip happens first, the card in spot 2 becomes mine.  Then, the cascade starts in spots 6 and 8, earning me spot 7 and 3 as well.  The cascade halts after that point, though, because there's nothing more to flip.

If the cascade happens first, then I get 6 and 8, plus the cascade through 7, 3, 2, and 1.  After the cascade, there's no flip on 2, since it was flipped in the cascade.

Poxnor

2012-08-30 05:46:54 pm

Moo! Flap! Hug!

What the heck? Watch the following:

At 1:05 of the video, a same is triggered (same-wall is active). But why the heck does the card in spot 5 flip (and cascade)? Obviously, there's something about these rules that I'm not understanding, at a more fundamental level...

Poxnor

2012-08-30 05:58:28 pm

Moo! Flap! Hug!

I'm wondering if the pseudo-code that implements it looks something like:

Code:

function playCard(c)
{
    Let cascade = Removing all elemental modifiers, is there a same, plus, or samewall triggered?
    if (cascade) {
        Let set = the 2, 3, 4, or 5 cards involved in the combo (including c)
        for each card x in set {
            Removing all elemental modifiers, check if x flips anything
            Add anything flipped to set
        }
    }
    Taking into account elemental modifiers, have c flip anything possible
}

If anyone has any counter-examples to that behaviour, or any thoughts about it, I would greatly appreciate it...

RaneofSOTN

2012-08-30 06:17:00 pm

I want off the ride....

I think same-wall has the same effect even if the numbersa ren't the same if the elemental takes action
So the 2-1 = 1 so it goes 1 -> 1 same wall and flips, as for why it might cascade... I dont know.

Poxnor

2012-08-30 06:20:06 pm

Moo! Flap! Hug!

I know for a fact that elementals aren't taken into account when determining if a same/plus/samewall is triggered. Just after 8:10 in that same video, a plus would have been triggered by the card played into spot 6 were elementals taken into account.

RaneofSOTN

Edit history:

RaneofSOTN: 2012-08-30 06:34:27 pm

RaneofSOTN: 2012-08-30 06:34:13 pm

2012-08-30 06:21:22 pm

I want off the ride....

I can only go off the knowledge i have from browser based triple triad I play and they seem to have that problem. I wonder if its a bug dealing with the Same wall? As it seems the flip shouldn't have happened but something with the graphic showing up of the same wall delayed the elemental action and cause the middle flip instead of a "push for nothing"

edit: NVM, it seems you have a same/wall that is forcing a lot of action off onto the middle card, and since it flips and a same/plus action happened, all cards that flip cascade. So because of the outer walls being treated as "A" value, the card he drops has an "A" facing the wall, and the bottom triggers cause 7 = 7.. so it would flip all cards around it cause of SAME/WALL ?...

so same/wall caused the cascade.

wiki for referece http://tripletriad.wikia.com/wiki/Triple_Triad:Triple_Triad_101#Same.2FPlus_Wall

Poxnor

Edit history:

Poxnor: 2012-08-30 06:42:05 pm

Poxnor: 2012-08-30 06:39:20 pm

2012-08-30 06:37:35 pm

Moo! Flap! Hug!

I didn't think a same/plus/samewall caused all the cards around the played card to flip (?). I thought it only (initially) flipped the cards that were involved in the same/plus (other cards could subsequently be flipped in the cascade, or combo as they call it in the wiki; but that card in spot 5 was flipped on the initial play, not as part of the cascade).

Edit, just to clarify: I see how the A and the 7 triggered a samewall. What I still don't get is what flipped the card in spot 5.

Edit 2: Come to think of it, I don't think I've ever tested the hypothesis that you were hinting at Rane -- that triggering a same/plus/samewall flips every card adjacent to the played card (and from there they cascade as possible). I'd find that really weird, but I'm not sure I've ever seen a counter-example to that.

RaneofSOTN

2012-08-30 06:43:26 pm

I want off the ride....

Regarding your Edit #2: on the triple triad games i've played through forums i've seen some of the "God bots" that use really weak cards, own the SHIT out of me because of that ruling with same wall, only reason i bring it up. I'm pretty sure it happens, its just same/wall/plus are not commonly played with regards to flipping more than 1-2 cards, and they are generally the ones you think of being used. I honestly never realized the outer walls were considered "A"

Poxnor

2012-08-30 06:47:21 pm

Moo! Flap! Hug!

Outer walls are only considered "A" if the samewall rule is active. If only "same" is active, but not "samewall," then the outer walls can't take part in a combo.

PJ

2012-08-30 07:02:26 pm

Is PJ

I remember getting raped hard by "blind, plus, same, combo, random" ruleset when I didn't know about manipulating rules. This card game is harder than the rest of the game if you don't know about eliminating rules.

Nitrodon

2012-08-30 07:25:54 pm

Good thing I was checking the AI code for you back then, since some of this is called from there.

Same rule (0009CB10-0009CD37 in RAM during Triple Triad):

Code:

Find the card that was played this turn.
Count the number of directions which fit the Same rule, including Same Wall if applicable.
   For each such direction with a card, set the Same rule flag for that card.
If fewer than two matches, stop.
For (each adjacent card) {
   If (card belongs to opponent and has a number less than or equal to the number facing it) {
      Flip card.
      Set success flag for Same rule.
      Set Same rule flag for played card.
}}

Plus rule (0009CD38-0009CF5B in RAM during Triple Triad):

Code:

Find the card that was played this turn.
For each adjacent card, compute the sum of the two numbers facing each other.
If no sum is represented twice, stop.
If two sums are represented twice, choose the sum which does NOT correspond to the card on the right.
For (each adjacent card with the appropriate sum) {
   Set Plus rule flag on adjacent card.
   If (card belongs to opponent) {
      Flip card.
      Set success flag for Plus rule.
      Set Plus rule flag for played card.
}}

The Same rule is handled first, followed by the Plus rule and the normal card flip. Combo is handled after all of these, but it seems that the AI doesn't check for combos at all, so I would need to search for it in the code.

Good thing I was checking the AI code for you back then, since some of this is called from there.

Same rule (0009CB10-0009CD37 in RAM during Triple Triad):
[code]Find the card that was played this turn.
Count the number of directions which fit the Same rule, including Same Wall if applicable.
   For each such direction with a card, set the Same rule flag for that card.
If fewer than two matches, stop.
For (each adjacent card) {
   If (card belongs to opponent and has a number less than or equal to the number facing it) {
      Flip card.
      Set success flag for Same rule.
      Set Same rule flag for played card.
}}[/code]

Plus rule (0009CD38-0009CF5B in RAM during Triple Triad):
[code]Find the card that was played this turn.
For each adjacent card, compute the sum of the two numbers facing each other.
If no sum is represented twice, stop.
If two sums are represented twice, choose the sum which does NOT correspond to the card on the right.
For (each adjacent card with the appropriate sum) {
   Set Plus rule flag on adjacent card.
   If (card belongs to opponent) {
      Flip card.
      Set success flag for Plus rule.
      Set Plus rule flag for played card.
}}[/code]

The Same rule is handled first, followed by the Plus rule and the normal card flip.  Combo is handled after all of these, but it seems that the AI doesn't check for combos at all, so I would need to search for it in the code.

Poxnor

2012-08-30 07:43:55 pm

Moo! Flap! Hug!

Thank you so much, Nitrodon! Smiley

Did someone take a hit to the head before they coded the same rule!? That doesn't even make sense...that success flag for the same rule? I hate to ask, but did you miss a line somewhere in the same rule code (something about flipping the cards that have the same flag set)?

I'm assuming the comparison between the two numbers facing numbers in the condition in the same code doesn't take into account elemental modifiers (which explains the mystery card flip in the video I posted).

Nitrodon

Edit history:

Nitrodon: 2012-08-30 07:58:11 pm

2012-08-30 07:58:01 pm

I didn't miss that line. Any cards with the Same flag set will have a number less than or equal to (specifically, equal to) the played card, and thus will be flipped. And yes, the Same and Plus routines completely ignore the Elemental rule.

I agree that the programmer in question must have taken a hit to the head. I did a bit of a double-take when I noticed it, but that's really what the code says. Doing it the correct way would actually take one fewer instruction.

Poxnor

2012-08-30 08:02:43 pm

Moo! Flap! Hug!

Quote from Nitrodon:

Any cards with the Same flag set will have a number less than or equal to (specifically, equal to) the played card, and thus will be flipped.

Ack, I missed the equality comparison (I misread it as "less than"). Long day, tired...sorry Smiley

Quote from Nitrodon:

And yes, the Same and Plus routines completely ignore the Elemental rule.

Okay, thanks!

Quote from Nitrodon:

I agree that the programmer in question must have taken a hit to the head. I did a bit of a double-take when I noticed it, but that's really what the code says. Doing it the correct way would actually take one fewer instruction.

Yup. It actually results in flips that shouldn't happen...

So about the only thing I'd need to know to actually expand the AI to play on a wider set of rules is what happens in the combo chains -- does the same code get recursively called (i.e., could you get same/plus happening during the combo phase), or is it doing something else...

Poxnor

Edit history:

Poxnor: 2012-08-30 08:10:39 pm

2012-08-30 08:06:23 pm

Moo! Flap! Hug!

It looks like Plus is implemented incorrectly as well. If you play a card into the middle spot and get two edges with a "plus sum" of 11, and two edges with a "plus sum" of 13...only one of the plusses will be activated instead of both.

Edit: lol, even worse, if two of the cards are yours and two of the cards are the opponent's, you might flip absolutely nothing, haha.

zoonel

2012-08-31 09:55:40 am

Go play spacechem !

Quote from Poxnor:

It looks like Plus is implemented incorrectly as well. If you play a card into the middle spot and get two edges with a "plus sum" of 11, and two edges with a "plus sum" of 13...only one of the plusses will be activated instead of both.

Edit: lol, even worse, if two of the cards are yours and two of the cards are the opponent's, you might flip absolutely nothing, haha.

Well, it is explained by Nitrodon code for the "plus rule"

Poxnor (out)

2012-08-31 10:03:58 am

Yeah, I know. I was making those observations based on the code, and how illogical it is...

zoonel

2012-08-31 03:25:28 pm

Go play spacechem !

Quote from Poxnor (out):

Yeah, I know. I was making those observations based on the code, and how illogical it is...

ok, I thought you had tested this ingame like the "same" rule.

Quick question for Nitrodon, you use the pc game or the emulated ps1 game for the code search ?

Poxnor

2012-08-31 05:47:23 pm

Moo! Flap! Hug!

He did it on the PSX code, but I'd be surprised if anyone rewrote the TT logic durin the PC port. It's just something that porters would have no reason to touch.

FFgamer86

2012-09-06 09:53:21 am

SDA Speedruns: 1

Couple new tricks discovered with the help of Tomato rice's run. Occupied Balamb now a minute quicker due to transitioning to hotel from the main street 9 times to have the guards unnerved by Raijin's late return causing Raijin to show up after talking to them, and Also an easier more consistent method to doing the Rinoa Skip that is more pushing through Watts Shoulder instead of walking Left between him and the door.