Board index 
Programs with AI or Rules Enforcement Magic: The Gathering - Duels of the Planeswalkers New MTG Cards and Decks (2010, 2012, 2013, 2014, 2015, Magic Duels) Magic Duels
Programs with AI or Rules Enforcement Magic: The Gathering - Duels of the Planeswalkers New MTG Cards and Decks (2010, 2012, 2013, 2014, 2015, Magic Duels) Magic Duels
Magic Duels tests and results
Moderator: CCGHQ Admins
Re: Magic Duels tests and results
by thefiremind » 08 Oct 2015, 08:44
Wizards said that they want to keep all the cards usable forever. I don't know how much we can trust them at this point, but I'd say it's more likely that they keep their word rather than not.petercal wrote:we get a card pool rotation once we hit the 1024 limit?
Yes, but maybe not in the way you're thinking of. If I understood it right, what the article says is that you need to store the card IDs inside the game states that will be evaluated by the AI. So, let's say that we have a game state where the AI needs to evaluate 10 cards (between hand and battlefield). Each card is represented by 10 bits, so the part of the game state pertaining the cards weighs 100 bits. If they extend the possible IDs they need to represent each card with more bits, so each game state will weigh more, and the CPU will need more time to make operations on it. I think it's a sensible problem because in Duels I see a noticeable increase in the AI's thinking time when a certain number of cards on the battlefield is reached. If I had seen such an increase in DotP2014 I would have checked if I had forgot to set AI_AVAILABILITY on some instant or ability, but in Duels it seems normal. I don't want to imagine the problem getting worse... the developers need to find a way to optimize the AI's evaluations.petercal wrote:The AI article basically implies the AI tree is correlated to the number of cards in system.
< Former DotP 2012/2013/2014 modder >
Currently busy with life...
Currently busy with life...
-
thefiremind - Programmer
- Posts: 3515
- Joined: 07 Nov 2011, 10:55
- Has thanked: 118 times
- Been thanked: 722 times
Re: Magic Duels tests and results
by spirolone » 09 Oct 2015, 15:46
I did some tests and I saw that it's possible to use id over 1024 even if game don't show those cards in deck builder; that is, we could actually create a community deck with 8192 cards!
Indeed, in profile, there are 13 bits for any cards in a deck. I think that, maybe, we could create a wad file with cards and use old Riiak's deck builder; then I can create a little java program to import decks...
Indeed, in profile, there are 13 bits for any cards in a deck. I think that, maybe, we could create a wad file with cards and use old Riiak's deck builder; then I can create a little java program to import decks...
- spirolone
- Programmer
- Posts: 190
- Joined: 31 Aug 2014, 23:14
- Has thanked: 7 times
- Been thanked: 107 times
Re: Magic Duels tests and results
by Xander9009 » 09 Oct 2015, 16:50
If you need any help that I can provide, let me know.
_______________________________
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
-
Xander9009 - Programmer
- Posts: 2905
- Joined: 29 Jun 2013, 07:44
- Location: Indiana, United States
- Has thanked: 121 times
- Been thanked: 445 times
Re: Magic Duels tests and results
by sweetLu » 14 Oct 2015, 21:35
Out of curiosity, where does the 8192 come from? That was someone's ID in 2014 and I just assumed it was a random number.spirolone wrote:I did some tests and I saw that it's possible to use id over 1024 even if game don't show those cards in deck builder; that is, we could actually create a community deck with 8192 cards!
Indeed, in profile, there are 13 bits for any cards in a deck. I think that, maybe, we could create a wad file with cards and use old Riiak's deck builder; then I can create a little java program to import decks...
Re: Magic Duels tests and results
by Xander9009 » 14 Oct 2015, 21:42
8192 is 2^13. So, in binary, it's 10000000000000 (1 and thirteen 0's). So, since there are 13 bits, there are 8192 combinations of 1's and 0's, making 8192 possible IDs.
_______________________________
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
-
Xander9009 - Programmer
- Posts: 2905
- Joined: 29 Jun 2013, 07:44
- Location: Indiana, United States
- Has thanked: 121 times
- Been thanked: 445 times
Re: Magic Duels tests and results
by zysron » 28 Oct 2015, 03:11
so i havent been on here for a while
but here is a thought being that i have not done any programming or if its possible but the card id's: the current ones, do they all have the same bit in the same spot? meaning is it possible that 1 bit may be a pointer to a wad file where the card resides? since i seriously doubt they plan of modding 1k+ cards when they get to that limit.
also pertaining to the game modes. currently it sucks but they advertised adding more gamemodes before they even put it out so patience... also if any of you are good coders this might be a way to figure out how to add your own gamemodes... also quick question
mods limit opponents still dont they?
but here is a thought being that i have not done any programming or if its possible but the card id's: the current ones, do they all have the same bit in the same spot? meaning is it possible that 1 bit may be a pointer to a wad file where the card resides? since i seriously doubt they plan of modding 1k+ cards when they get to that limit.
also pertaining to the game modes. currently it sucks but they advertised adding more gamemodes before they even put it out so patience... also if any of you are good coders this might be a way to figure out how to add your own gamemodes... also quick question
mods limit opponents still dont they?
Re: Magic Duels tests and results
by Xander9009 » 19 Nov 2015, 20:33
Ok, so does anyone know what the final verdict was for the maximum number of cards? Spirolone mentioned 8000 being do-able. Is that the result? If so, which cards do we want to do? Currently we have 50% more than that in DotP 2014. So, we can only convert about 2/3rds of the collection. Which ones should make the cut? All mythic rares and rares? Do we judge each card based on how many decks it's used in? Rank them based on that, and simply stop when we run out of room?
_______________________________
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
-
Xander9009 - Programmer
- Posts: 2905
- Joined: 29 Jun 2013, 07:44
- Location: Indiana, United States
- Has thanked: 121 times
- Been thanked: 445 times
Re: Magic Duels tests and results
by Xander9009 » 19 Nov 2015, 22:56
Alright. So, I made a quick tool to list out all cards with a unique number based on the deck they're from to the clipboard which could be pasted elsewhere. Using this with Notepad++ and Excel, I narrowed it down to this list. This list does NOT include base-game decks. Since there are fewer than 4000 cards actually in use (there are 3058), I find it reasonable to convert those cards, all base-game cards, and all cards from previous games. Does this sound reasonable to others? (Note that the decks I checked are ALL decks in the CW Decks folder, which includes almost all of the mods here, including sumomole's thefiremind's, RiiakShiNal's, and of course anyone like myself and Kithkin who only released decks via the CW.)
For the curious, here is the list. Due to the post size limit, I had to remove the "1 - "'s from before the cards at the end. If there's no number, it's used once.
For the curious, here is the list. Due to the post size limit, I had to remove the "1 - "'s from before the cards at the end. If there's no number, it's used once.
- Card Use Count | Open
_______________________________
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
-
Xander9009 - Programmer
- Posts: 2905
- Joined: 29 Jun 2013, 07:44
- Location: Indiana, United States
- Has thanked: 121 times
- Been thanked: 445 times
Re: Magic Duels tests and results
by thefiremind » 20 Nov 2015, 00:47
There's a greater problem at hand: I can't seem to overwrite the card pools anymore. If I make the usual ZED file with my extened shoebox pool I see only BFZ cards in my collection, as if the game decided that the other pools were damaged. If I try to change only 1 card inside the new BFZ shoebox, my new pool is just ignored, and I still see the original card. Could it be because I'm still building the ZED with the encrypted data from DATA_000, while DATA_001's folders are set to a higher order priority inside .METADATA\Header.xml? If that's the case, I need to wait for spirolone to extract the encrypted data from DATA_001 and make a DATA_001.bin file.
Actually, since the problem may rise again with the next expansion, I'd suggest spirolone to make something that doesn't need his intervention anymore: either a "ZedToBin" program that allows anyone to extract the encrypted part, or a modification to the ZED building process that allows to use ZED files directly instead of having to use his BIN extractions.
Actually, since the problem may rise again with the next expansion, I'd suggest spirolone to make something that doesn't need his intervention anymore: either a "ZedToBin" program that allows anyone to extract the encrypted part, or a modification to the ZED building process that allows to use ZED files directly instead of having to use his BIN extractions.
< Former DotP 2012/2013/2014 modder >
Currently busy with life...
Currently busy with life...
-
thefiremind - Programmer
- Posts: 3515
- Joined: 07 Nov 2011, 10:55
- Has thanked: 118 times
- Been thanked: 722 times
Re: Magic Duels tests and results
by spirolone » 21 Nov 2015, 05:59
Here is a new Data.bin obtained from DATA_001.ZED. I attached also a lighty modified CreateZip.jar that uses DATA_001 as default source folder and that allows to set it as argument. Remember that now you have to use DATA_001 as name of folder for your content.thefiremind wrote:There's a greater problem at hand: I can't seem to overwrite the card pools anymore.
...
Then I created a new jar, ZedToDataBin, following thefiremind suggestion. It creates new Data.bin file from a zed file; you can set name as argument and default is DATA_001 (".ZED" is added by program).
Actually I'm doing tests on new version, but I wanted to upload those files quickly to let others do other tests...
- Attachments
-
Zed.zip- (59.31 KiB) Downloaded 340 times
- spirolone
- Programmer
- Posts: 190
- Joined: 31 Aug 2014, 23:14
- Has thanked: 7 times
- Been thanked: 107 times
Re: Magic Duels tests and results
by thefiremind » 21 Nov 2015, 10:47
Thanks spirolone! However, it turned out it may have been my fault because I forgot a " in the new pool. 
I'll report more results as soon as I can.
EDIT: Confirmed, it was my fault. At the moment I'm trying to code bestow by using awaken as a template. Assuming I can make it work, there's at least one approximation to take into account: according to the rules, cards like AEther Storm should still allow to cast bestow cards as Auras, but in Duels that will be impossible I'm afraid.
EDIT 2: Here's my bestow mechanic for Magic Duels in the form of Nyxborn Shieldmate:
Note: I changed my TFM_RemoveCardType function so that it doesn't call ClearAllOfType, because type changes should be made on layer 4A while subtype changes should be made on layer 4B. Ephara was working even by doing everything on 4B, but I want to adhere to the rules so that we don't find unexpected bugs in the future.
I'll report more results as soon as I can.EDIT: Confirmed, it was my fault. At the moment I'm trying to code bestow by using awaken as a template. Assuming I can make it work, there's at least one approximation to take into account: according to the rules, cards like AEther Storm should still allow to cast bestow cards as Auras, but in Duels that will be impossible I'm afraid.
EDIT 2: Here's my bestow mechanic for Magic Duels in the form of Nyxborn Shieldmate:
- Nyxborn Shieldmate (seems to work) | Open
Note: I changed my TFM_RemoveCardType function so that it doesn't call ClearAllOfType, because type changes should be made on layer 4A while subtype changes should be made on layer 4B. Ephara was working even by doing everything on 4B, but I want to adhere to the rules so that we don't find unexpected bugs in the future.
Last edited by thefiremind on 21 Nov 2015, 19:29, edited 1 time in total.
< Former DotP 2012/2013/2014 modder >
Currently busy with life...
Currently busy with life...
-
thefiremind - Programmer
- Posts: 3515
- Joined: 07 Nov 2011, 10:55
- Has thanked: 118 times
- Been thanked: 722 times
Re: Magic Duels tests and results
by Xander9009 » 21 Nov 2015, 18:25
TFM, that first edit was a very confusing lol. I'm pretty sure every instance of bolster was supposed to be bestow. IT's good to hear that it is working after all, though. 
This implies we can convert at least the cards.
@All: I've decided that since only 3220 cards have actually been used, that's what I'll focus on. So, I have a question. Given that the card cannot be viewed in the deck builder once we surpass 1024, I take it this means we need the deck builder updated to output to the Magic Duels deck format, and the cards need added to a zed file. The question: what changes need to be made to a card to make it work with that? Just the changes needed for a simple card, not the functions changing or the constants being renamed/removed and things like that. Looking at the (mostly?) working Nyxborn Shieldmate, it doesn't have any changes that would absolutely be needed on ever card (so, no card structure changes). The biggest change I see (and it was suggested the previous way still worked?) is that the text permanent strings can be placed directly on the card. If the previous method still works, the obviously that would be better for any text used on more than 2 or 3 cards, if only to avoid extra size. Does that all sound right?
This implies we can convert at least the cards.@All: I've decided that since only 3220 cards have actually been used, that's what I'll focus on. So, I have a question. Given that the card cannot be viewed in the deck builder once we surpass 1024, I take it this means we need the deck builder updated to output to the Magic Duels deck format, and the cards need added to a zed file. The question: what changes need to be made to a card to make it work with that? Just the changes needed for a simple card, not the functions changing or the constants being renamed/removed and things like that. Looking at the (mostly?) working Nyxborn Shieldmate, it doesn't have any changes that would absolutely be needed on ever card (so, no card structure changes). The biggest change I see (and it was suggested the previous way still worked?) is that the text permanent strings can be placed directly on the card. If the previous method still works, the obviously that would be better for any text used on more than 2 or 3 cards, if only to avoid extra size. Does that all sound right?
_______________________________
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
Community Wad - Community Wad Website - How to Help and Report Bugs
Discord: discord.gg/4AXvHzW
-
Xander9009 - Programmer
- Posts: 2905
- Joined: 29 Jun 2013, 07:44
- Location: Indiana, United States
- Has thanked: 121 times
- Been thanked: 445 times
Re: Magic Duels tests and results
by thefiremind » 21 Nov 2015, 19:49
I'm not sure what I was thinking when I wrote that. Fixed for clarity.Xander9009 wrote:TFM, that first edit was a very confusing lol. I'm pretty sure every instance of bolster was supposed to be bestow.
My take on bestow is personal, I took a look at NeoAnderson's code only to find inspiration on how to end the effect (I ended up using a LinkedDC flag as he did).Xander9009 wrote:IT's good to hear that it is working after all, though.
For those who stopped at DotP2014, some of the changes were made in the transition to DotP2015 and can be found here. Be aware they don't all carry over (e.g. Planeswalkers now work) and there are some I haven't had the chance to re-test, but I'm testing now. The structure is still the same anyway, and yes, Text_Permanent XML files still work. There is a new SET_ID block in the new cards, but it's about the expansion symbol (not needed for functionality) and I'm not even sure about how it works: my cards don't have SET_ID and I guess that the game gives the symbol based on the pool the cards belong to when it's missing.Xander9009 wrote:what changes need to be made to a card to make it work with that? Just the changes needed for a simple card, not the functions changing or the constants being renamed/removed and things like that. Looking at the (mostly?) working Nyxborn Shieldmate, it doesn't have any changes that would absolutely be needed on ever card (so, no card structure changes). The biggest change I see (and it was suggested the previous way still worked?) is that the text permanent strings can be placed directly on the card. If the previous method still works, the obviously that would be better for any text used on more than 2 or 3 cards, if only to avoid extra size.
EDIT: Despite having a mana pool now, the BECAME_TAPPED_FOR_MANA trigger is still disabled. I tried by both auto-tapping and manually tapping a Zhur-Taa Druid.
EDIT 2: SetBlockController is still broken.
< Former DotP 2012/2013/2014 modder >
Currently busy with life...
Currently busy with life...
-
thefiremind - Programmer
- Posts: 3515
- Joined: 07 Nov 2011, 10:55
- Has thanked: 118 times
- Been thanked: 722 times
Re: Magic Duels tests and results
by spirolone » 22 Nov 2015, 00:40
I hoped to obtain new informations about cards limit with new version, but I didn't: I don't know how to unlock cards with ID over 1023 so that we can use them in game deck builder, and I don't know if there is a way to use a blank byte that is reserved for any card in decks inside profile file, and that, maybe, could expand 8191 cards limit...
Now, I successfully used an external deck editor to create decks and imported them in profile. I used a Java program that reads output xml file of deck editor and compares any card name with a cardpool file to obtain ID used in game.
I used Magic Assistant as deck builder: what do you think about it? If we choise a standard format for decks, I can upload a jar to import a deck in profile so that we can use it in game...
"Imported" decks can be edited by game deck builder: it doesn't show with zero as "availability" but if it is negative, it show that card! That is, if you have in your deck 1 card you don't own, you can see that card in available ones with availability "x-1" and you can add other instances of that card!
Now, I successfully used an external deck editor to create decks and imported them in profile. I used a Java program that reads output xml file of deck editor and compares any card name with a cardpool file to obtain ID used in game.
I used Magic Assistant as deck builder: what do you think about it? If we choise a standard format for decks, I can upload a jar to import a deck in profile so that we can use it in game...
"Imported" decks can be edited by game deck builder: it doesn't show with zero as "availability" but if it is negative, it show that card! That is, if you have in your deck 1 card you don't own, you can see that card in available ones with availability "x-1" and you can add other instances of that card!
- spirolone
- Programmer
- Posts: 190
- Joined: 31 Aug 2014, 23:14
- Has thanked: 7 times
- Been thanked: 107 times
Who is online
Users browsing this forum: No registered users and 0 guests