[Request/Question] Triple Triad Plugin

[flamepyro]

So my Request/Question is about TT how feasible is it to get a plugin to farm a TT over and over again?

Also do we have Tags in the bot for these?

Thanks for any help out there.

 

[DBG]

The logic for playing TT is not in place yet (afaik, corrections are always welcomed as I'm not at my desktop atm - I just don't remember seeing any new hooks added), but I cannot speak for what's in development. From a programming standpoint it would not be hard to code just for (I already have code for free and plus mode), the trouble is any plugin I might be able to create right now would basically just be able to only beat "Triple Triad Master" and "King Elmer III" 100% of the time as a lot depends on what cards/decks you have. The Golden Saucer and its activities are obviously very popular and there are people with much greater technical minds than mine looking at tackling all of these games (but ofc I'm looking into it myself as well since my first WIP plugin of having alternative mounts was made null and void by the man with the master plan ).

 

[flamepyro]

Thats kinda what im looking for as i want something that can farm MGP well and it seems TT is the best at it with the least amount of risk.

Basicly ill get the cards needed and would have the plugin just run its thing with a blank profile

 

[kagamihiiragi17]

The logic for playing TT is not in place yet (afaik, corrections are always welcomed as I'm not at my desktop atm - I just don't remember seeing any new hooks added), but I cannot speak for what's in development. From a programming standpoint it would not be hard to code just for (I already have code for free and plus mode), the trouble is any plugin I might be able to create right now would basically just be able to only beat "Triple Triad Master" and "King Elmer III" 100% of the time as a lot depends on what cards/decks you have. The Golden Saucer and its activities are obviously very popular and there are people with much greater technical minds than mine looking at tackling all of these games (but ofc I'm looking into it myself as well since my first WIP plugin of having alternative mounts was made null and void by the man with the master plan ).

What do you mean you have code for Free and Plus mode?

 

[iyake]

I found it easier to just script out the plays a particular opponent plays and the responses to them. They have a fairly limited set of plays they'll make, so you could potentially exhaust all possible plays to guarantee 100% winrate.

 

[kagamihiiragi17]

I found it easier to just script out the plays a particular opponent plays and the responses to them. They have a fairly limited set of plays they'll make, so you could potentially exhaust all possible plays to guarantee 100% winrate.

Not the best strategy for a bot though, if you have enough logic then you should be able to beat almost any opponent without scripting them all.

 

[Tinytox]

What do you mean you have code for Free and Plus mode?

I'm sure that means..
Having logic for different rule sets, such as Plus, Same, etc.
Though.. there's a LOT of rules, including ones that can really screw you over, Random, Chaos, etc.

Depending on the opponent and the cards you draw (since AI isn't limited to certain card rarities..) and the rule set, you can basically draw a hand that will guarantee a loss without even having to play. This is reinforced by the "random" ruleset, while of course, the wonderful system SE set up, the AI's card bank can be like, 7/8 cards.. so it "randomly" picks for 7/8 ridiculously good cards.

Point is the system is pretty unfair and half the time you're forced to try even just to get a good hand dealt. It gets super tedious.

Also a side note, Many NPC's adhere to regional rules which change often (weekly reset, I believe.), That being the case, it would often 'lock' an opponent in terms of being able to bot them/play them effectively (even legit)... Because the wrong rule sets can really be a slap in the face.

 

[kagamihiiragi17]

I'm sure that means..
Having logic for different rule sets, such as Plus, Same, etc.
Though.. there's a LOT of rules, including ones that can really screw you over, Random, Chaos, etc.

Depending on the opponent and the cards you draw (since AI isn't limited to certain card rarities..) and the rule set, you can basically draw a hand that will guarantee a loss without even having to play. This is reinforced by the "random" ruleset, while of course, the wonderful system SE set up, the AI's card bank can be like, 7/8 cards.. so it "randomly" picks for 7/8 ridiculously good cards.

Point is the system is pretty unfair and half the time you're forced to try even just to get a good hand dealt. It gets super tedious.

Also a side note, Many NPC's adhere to regional rules which change often (weekly reset, I believe.), That being the case, it would often 'lock' an opponent in terms of being able to bot them/play them effectively (even legit)... Because the wrong rule sets can really be a slap in the face.

I understood what he meant, I suppose my question was more "Why do you have code for those rulesets when there's no API to support them?"

I'm trying to think of how to best tackle the logic problem for TT, I'm only just starting FF8 and I've only played a couple matches in FFXIV. Say for example we wanted to automate King Elmer III or w/e his name is to farm his cards and his MGP, do you have any strategies for the Plus rule? My thought would just be to iterate through every possible move, figure out what move takes the most cards into our control and play that, or if there are no moves to capture cards then play the card where the sides which are adjacent to walls or placed cards totals to the smallest number (for safety).

 

[kagepande]

With King Elmer III, you can build a specific deck(pretty much all starter cards) and you try and force him to play into a certain way so that you can use the plus rule to take back 66% or more of the board in one go. There is a set number of ways that he will play his cards based off how you play yours, once you figure these out you can get a near 100% win rate against him by hand, and if you don't win you can at least tie, which still earns you MGP.

 

[kagamihiiragi17]

With King Elmer III, you can build a specific deck(pretty much all starter cards) and you try and force him to play into a certain way so that you can use the plus rule to take back 66% or more of the board in one go. There is a set number of ways that he will play his cards based off how you play yours, once you figure these out you can get a near 100% win rate against him by hand, and if you don't win you can at least tie, which still earns you MGP.

Do you happen to have that information?

 

[kagepande]

In my head yes, Im not 100% on all his possible plays, I'm not the best Triple Triad player so I don't know how to beat every single way he might play his cards. But this is this Youtube video that I used right when it was released and have had really high success rates.

https://www.youtube.com/watch?v=Uu-jRcMQ_10

 

[kagamihiiragi17]

In my head yes, Im not 100% on all his possible plays, I'm not the best Triple Triad player so I don't know how to beat every single way he might play his cards. But this is this Youtube video that I used right when it was released and have had really high success rates.

https://www.youtube.com/watch?v=Uu-jRcMQ_10

Thanks, I'll check it out

 

[Tinytox]

Do you happen to have that information?

I can't remember what setup I used but I eventually found that there's a very solid method to abusing elmer.
I believe I used... Sabotender, Bomb, Coerl, Chocobo, and Ixali (Really recommend Coblyn instead of Coerl though)

I used a starter deck and went to town, Basically your goal is to bait him into moving in certain spots while leaving the (usually center, although it's doable in other spots).. open

If you plan to bait him on the center position, then you most importantly need to have him not place any cards there, that's honestly the tricky part, and sometimes he does it anyway.. Even if it wouldn't really benefit him.

I found that I had the highest success rate playing my cards on the sides, (not corners) And near the end (depends who started), I would plus against cards that I originally placed down. Now.. A -LOT- of people recommend the godbert card and some others for this particular setup, but I didn't and still don't have it. (Once I got bahamut I booked, that's all I really wanted.)

There's a reddit thread that goes over a similar strategy, but requires the Minfilia card... With this in mind, we wind up in that oh-so-common "TT is impossible to get started in." problem, that actually plagues a lot of normal players. It seems for many people it's impossible to get 1 very powerful card, thus they can't continue, because their deck is too weak to get any better cards, and of course, thanks to the rarity limitations, having several "decent" cards (2-3 star) won't help you until you have 30+/60+ cards.
Triple Triad - Grinding King Elmer III : ffxiv

I can't really remember the deck I used, but the important part about the deck you use is really only that you can plus the cards that you placed, (The goal is to plus off of one of your own and one of his cards if possible, as the resulting "Combo" will help solidify a win, otherwise it can be a real shit toss.)

One last little note, Thanks to Square being retarded, there's glaring problems with Triple Triad that make the average FF8 fanboy instead.. The most obvious of those being the rarity rules preventing you from ever having more than 1 awesome card. That makes ever 'being the best' or, in FF8, having all player cards, etc, completely impossible.

I was going to try to become a TT master in FF14, then I farmed Dragons Neck 80 times and never got the Terra card, I cried myself to sleep and gave up after that. Maybe I'll try again someday.

Little side note.. Now that I'm thinking about it, I seem to remember not using the center spot to PLUS him, I feel like it was the middle left slot, but, can't really remember, I seem to remember that playing in the corners is usually a problem, though. And I also remember that the style of play changed quite drastically depending on who went first.

 

[Hbk]

In my head yes, Im not 100% on all his possible plays, I'm not the best Triple Triad player so I don't know how to beat every single way he might play his cards. But this is this Youtube video that I used right when it was released and have had really high success rates.

https://www.youtube.com/watch?v=Uu-jRcMQ_10

While having logic to win TT would be ideal, I think that most people would agree that having a plugin to beat just one NPC is what they're looking for. From my experience the NPCs have a pretty limited amount of moves that they will make, depending on what you play and where. Even if the bot was able to win ~60% of its matches, it would be incredibly valuable for anyone farming MGP. Can't wait to see this in the API or to see a plugin made.

 

[Tinytox]

While having logic to win TT would be ideal, I think that most people would agree that having a plugin to beat just one NPC is what they're looking for. From my experience the NPCs have a pretty limited amount of moves that they will make, depending on what you play and where. Even if the bot was able to win ~60% of its matches, it would be incredibly valuable for anyone farming MGP. Can't wait to see this in the API or to see a plugin made.

Even just a plugin that would allow users to make their own, custom responses would be viably usable.. Hell, even if they didn't support any form of checks, using a single strategy against the same opponent (allow for two different methods based on who goes first, though!) would be enough to win like 30% of games, most likely. Depends on the opponent they want to win a specific card from, but, say that opponent is.. oh.. i don't know, Elmer for instance (easy target!), 30% win rate would probably keep you MGP on par at least, and if that's "REALLY" an issue for you, low and behold, there's a MGP farming plugin for chocobo races already available.

If we REALLY had to be minimalist, a simple way of scripting a match would probably be enough, especially because the AI seems to have a pretty limited array of variance to certain moves, most of the time it seems that in response to a specific move, the whole board isn't available to them. With this in mind, scripting a match should be fairly reasonable, right?

I suppose it could be a fickle thing, and that full blown logic should be added on, to do the job correctly, however, there's a lot of crap to account for. Ultimately, I think it's safe to say that it would be more work than it's worth to do anything more than the minimal work required, say, if the opponent usually plays one way, and they play any way other than how we expect them to, then RB could default to "slap all my cards down" aka "I give up.", and begins the rematch anew.. while it would be strikingly less effective, I think it's viable, though there are a loooooooot of viable options, and TT does boil down to a lot of math with some randomization involved, which means it is all calculable, the hard part is really just adjusting for the npc's DERP moments.

Really, I think there's only a few cards most people want to have the option to bot, probably.. Bahamut, Minfilia, and I'm sure there are a few others, too.

I'm fairly confident many of the creators here could do it, but to do it 'right' is a daunting task (See: SAO Suite)

 

[kagepande]

Its not for a bot, but I have found a few AHK scripts that can get a 80% or so win rate, 18% draw rate and 2% loss rate, roughly. Its possible its just a matter of having support for it all

 

[Snipi]

Its not for a bot, but I have found a few AHK scripts that can get a 80% or so win rate, 18% draw rate and 2% loss rate, roughly. Its possible its just a matter of having support for it all

Can you link this AHK script pls ? I've found one but he isn't finished...

 

[ryudo6850]

Its not for a bot, but I have found a few AHK scripts that can get a 80% or so win rate, 18% draw rate and 2% loss rate, roughly. Its possible its just a matter of having support for it all

If it's the one by brooklynbase... I think it has a much higher loss rate. Thus I created one that takes into account for more scenarios, a different deck (still need 60 cards). Instead of verifying your cards as the method of placing; mine checks certain open spots based off patterns I remember when farming him at 100% win rate. However automation isn't perfect and to make it perfect I would need to create something far more elaborate then I intend to do so. Still with this one you SHOULD never lose and I have yet to do so.

I think this script has like 80-90% win rate and a 5-15% draw rate or something.

Now I'm not sure what to do it with it yet; I might find one person who would test, add, and improve upon it.(If you are interested in being that person send me a PM.)


As of now there isn't a public release until I feel it would be worthy and be near perfect. Plus right now people will have to still enter their own pixel colors and the position of your cards + panels. All-in-all this means that the people who probably want it the most couldn't use it. Those who would know how to use it likely have BB's script and changed it for themselves.