Magic Duels tests and results

[thefiremind]

so what now? do we just abandon DotP2014 and DotP2015? I am getting a little confused on where should I focus D:

I'm always only interested in the more recent content, but everyone should do as they want. Moreover, it seems this got almost no interest, so I guess the majority wants to keep DotP2014 as the preferred modding platform. I can understand why: we have few information about modding Duels (of course it won't increase if nobody bothers), and Duels doesn't allow to make the AI use our decks. I hope that, at some point, the pros will surpass the cons... we'll see what's new when Battle for Zendikar gets added. I stopped my tests until that time comes.

[petercal]

Apparently, we will be getting additional "D" cards to the starter box (free cards) when the BFZ expansion hits. According to the Wizards answer on reddit, the new "D" cards appear after "you complete the new story campaign or you purchase your first pack of Battle for Zendikar cards".

Perhaps it is a good idea to preserve a copy of your Duels game now for comparison before and after so we see the activation changes?

[Chakan]

I for one, will be sticking with Magic 2014. Not only is it the known entity when it comes to modding, it's a pretty powerful platform for Magic as a whole. Almost anything is possible with the engine and the fundamentals of the game are good. You can have the AI use decks you want them to use, there's Two Headed Giant (better than Duels, and of course 2015 since it didn't have it), better multiplayer lobbies, you can play against AI decks in Two Headed Giant, and 2014 doesn't have any sort of limits. The game is still fun, the AI is smarter than you'd expect it to be... There's a lot more going on for 2014 than I think Duels will ever have going for it.

Duels is a decent game, but I don't think it will ever be modded properly or as easily as 2014 is. Even if it's possible, the game has a lot of limits to it.

[GrovyleXShinyCelebi]

I'm the only one that's modding Magic: Duels Origins...

Honestly, the only real restrictions I see are the 1024-card limit and maybe the deck builder which has lost some functionality since 2015 such as card sorting. If there's a way to either extend or get around that limit Origins would be an excellent platform to mod. We already have ways to open and pack ZED files at will, there's planeswalkers and dual lands, 2HG is back (and playable against the AI if you have a USB controller), you can make whatever deck you want, etc.

It does appear to be possible to get the AI to use your cards though, although I haven't tested it yet. Under the ExploreDecks directory in DATA_000 there appear to be a list of all the explore decks available, which interestingly enough is not generated randomly but are actually presets (there are 1680 different presets for each difficulty). You're probably able to add new presets or change them to throw in some modded cards.

[Xander9009]

The 1024 card limit is my main concern. Navigating around that would be the single most important thing for me to consider switching.

[thefiremind]

At some point, that limit will become too restrictive for the developers as well, and they'll need to overcome it. Let's hope they make the necessary adjustments before that happens (I don't know how many expansions will take to reach 1024). I think that their main problem will be to reorganize the profile.

[Xander9009]

What I'm worried about is that they'll increase it to 2048 when they need to. And then when they need to (assuming it ever gets to that point), increase it only to 3072; adding blocks of 1024 rather than doubling or removing it.

[Zambooo]

I think Magic Duels is the way too: we can modify the cost for activated ability, dual-trilands and mana of any color (who were also in 2015), planeswalkers, dual cards, "cannot be blocked by more than one creature" and hoping they bring it from 2015, "Convoke" (and I am sure I didn't list all the feature 2014 doesn't have). Also, I am confident WotC will itself raise the 1024 card limit with new updates.
Moreover we need to consider that Origins will be a persistent client, which means that we will get directly from the devs new cards and ability with every new expansion! Just think, if they sooner or later bring back with an expansion Morph or Manifest we will just get the ability coded.

[GrovyleXShinyCelebi]

What I'm worried about is that they'll increase it to 2048 when they need to. And then when they need to (assuming it ever gets to that point), increase it only to 3072; adding blocks of 1024 rather than doubling or removing it.

I think the main reason why there is a hard 1024 limit is because the card ID's are 10 bits long (2^10 = 1024). Honestly, if we assume about 162 cards per expansion (that's how many cards there were here), and that this is a large expansion- smaller ones will have less cards, we'll need at least five sets before they add another bit to the card ID's, so we'll need a way around the limit if we want serious modding within the year.

Oh yeah, I checked convoke from imported DOTP 2015 cards and it doesn't work in Origins. If I recall correctly the mana wasn't being subtracted properly (don't quote me on this). Not surprised- it was a really complicated mechanic. Hardcoded too...

Annihilator works now with a small change: with the "ATTACKING" trigger "TriggerPlayer()" is now "EffectSource():GetDefendingPlayer()".

[Karmabot]

The only question should be what the newest version has to Offer that the old Duels one doesnt.

Personally im an advocate for the new version since as i understand it will not be invalidated every year but constantly updated.
Meaning that if we keep our Code properly encapsulated it will continue working on the up to date platform that after a while will be more feature rich and stable.

Im wondering about the 1024 Card limit though, since most of the game is running in lua we should be able to freely mod it by having the Engine run our scripts.

Ill check it out

[Xander9009]

The only question should be what the newest version has to Offer that the old Duels one doesnt.

Personally im an advocate for the new version since as i understand it will not be invalidated every year but constantly updated.
Meaning that if we keep our Code properly encapsulated it will continue working on the up to date platform that after a while will be more feature rich and stable.

Im wondering about the 1024 Card limit though, since most of the game is running in lua we should be able to freely mod it by having the Engine run our scripts.

Ill check it out

The limit is the only thing I consider a serious con. I don't like the fact that I can't force my own decks to be used, I don't like that I can't play various game formats, I don't like the built in deck builder. There are many things I dislike about Magic Duels. But I don't personally consider any of them to be worth not moving the modding scene there except the card limit. If we can get around that by forcing the engine to load more and treat the profile differently to mirror that larger load, then I would definitely consider the move to that as a modding community to be the best way forward. But until that limit is gone, the most we could do is have one mod of ~700(?) cards active at a time. This means the mods would need to also have the profiles included (or some patcher to include the decks in the current profile), making using modded decks alongside an individual user's custom decks rather difficult.

[petercal]

Wizards already announced that full-art lands will not be in the BFZ update, disappointing many people.

So I was wondering how much effort it would be to implement full-art lands as a replacement for the five basic lands. Is it a straight graphic file replacement sufficient or is a new frame graphic required as well (can it be a drop in replacement)?

[thefiremind]

Is it a straight graphic file replacement sufficient or is a new frame graphic required as well (can it be a drop in replacement)?

Frame choice has been limited since DotP2013, we can't just add a new frame and make the game use it (unless the executable already contains its name, as for the Nyx frames for double-colored cards), otherwise the problems I listed with transform cards and Planeswalkers wouldn't exist. Actually, full-art lands would have been a problem even before DotP2013, because the card type text would still be put in the middle and it wouldn't be aesthetically pleasant at all.

[petercal]

Is it a straight graphic file replacement sufficient or is a new frame graphic required as well (can it be a drop in replacement)?

Frame choice has been limited since DotP2013, we can't just add a new frame and make the game use it (unless the executable already contains its name, as for the Nyx frames for double-colored cards), otherwise the problems I listed with transform cards and Planeswalkers wouldn't exist. Actually, full-art lands would have been a problem even before DotP2013, because the card type text would still be put in the middle and it wouldn't be aesthetically pleasant at all.

Bummer. I take it then that the BFZ frames have not been added to the executable yet. Even so, I understand now that the likely overriding problem is how the card type text is displayed in the middle of the card. I assume re-coding (more trouble than it is worth?) is necessary to provide display flexibility to accommodate full-art cards.

[petercal]

I respectfully seek some clarification from modders that may know how Duels engine could work.

On another discussion I noted that unless Stainless changes the way the Duels engine currently uses CARD IDs (10-bit number), the Duels card pool will max out at 2^10 = 1024 cards.

To which I got this interesting reply:

If they're using a 10-bit card ID, I can almost guarantee that they're packing the 10 ID bits and six status flags (like "tapped", "face-up/face-down", "front-face/back-face") into a 16-bit word. When the AI is making copies of the engine state to try different possibilities in, the size of each state has a big influence on performance. Being able to fit one more card in each CPU cache line could make the difference between the AI noticing a possibility and timing out before it tries that line.

This got me thinking about the possible difficulty, & thus likelihood, of increasing the upper limit of CARD ID. The AI article basically implies the AI tree is correlated to the number of cards in system. The more cards, the larger number of branches to consider. Thus, we have a trade off in terms of AI performance versus card pool size. However, this begs the question how well does DotP 2014 perform with all the community mods?

Ultimately, my question is: if the CARD ID upper limit is linked to the way AI works (or am I misunderstanding the whole thing?), how likely can this limit be increases. Looking at problem from the non-modder perspective, the question is which is more likely: 1) we see an increase in card pool beyond 1024, or 2) we get a card pool rotation once we hit the 1024 limit?