Getting Fallen Empires draft to work

[lujo]

Right, so here's the provisory Merseine code I'm using for testing, with the compromises until there's proper tools to build it right, as instructed by friarsol:

Provisory Mersiene Code | Open

Code: Select all

Name:Merseine
ManaCost:2 U U
Types:Enchantment Aura
K:Enchant creature
A:SP$ Attach | Cost$ 2 U U | ValidTgts$ Creature | AILogic$ KeepTapped
K:etbCounter:NET:3
S:Mode$ Continuous | Affected$ Creature.EnchantedBy | AddHiddenKeyword$ CARDNAME doesn't untap during your untap step. | CheckSVar$ X | Description$ Enchanted creature doesn't untap during its controller's untap step if Merseine has a net counter on it.
SVar:X:Count$CardCounters.NET
A:AB$ RemoveCounter | AnyPlayer$ True | Cost$ 4 | CounterType$ NET | CounterNum$ 1 | SpellDescription$ Remove a net counter from CARDNAME.
SVar:Picture:http://www.wizards.com/global/images/magic/general/merseine.jpg
Oracle:Enchant creature
Merseine enters the battlefield with three net counters on it.
Enchanted creature doesn't untap during its controller's untap step if Merseine has a net counter on it.
Pay enchanted creature's mana cost: Remove a net counter from Merseine. Any player may activate this ability, but only if he or she controls the enchanted creature.

-
-Seems to be working as well as it should, having it feels incredibly good.
-
And the reason I've been putting off posting decks up is that I want to see if a few things can be resolved before that. For instance I've created an "Unsorted Treasure Token" picture to use in place of Raiding Party, with no application but being able to be sold in the quest shop and making the draft have the correct number of cards of all rarities and all print runs working and did a few pods with it, it was nice.

But while exploring eclipse and the code I got stuck on the idea of making it a flip card which flips between the fabulous treasure art and the original card art with something like this as the "Treasure" side:




I'm not stuck on that idea, I can just make a regular picture for it and have Forge use it instead of the intended art for the unimplemented card and implement the card as a blank, manacostless one which the AI ignores, but I was curious if I could find a way to do it on my own.

Best I've got is for the forge to read the card, but have no clue what set it's from and not read either of the arts.

Been doing some other stuff too, gotta find a way to see how to get a few deckbuilding labels which limit the desired number of cards after which the AI doesn't prioritize them in draft or deckbuilding - the decks look better but certain AI's tend to overdraft stuff due to the "one big decklist" problem, so one guy will get 6 copies of one common the other of another and so on.

Also waiting on potential confirmation on the fixes for Orcish Captain AI tweak and the Night Soil interface bug solutions which I reported in other threads and making a ton of notes.

Oh, and I wanted to put up a few short videos, because it's easier to explain and showcase stuff with words and pictures rather than textwalls, but I had to put that off until I went and scored some more RAM for my laptop and get a new non-humming cooler fan, RAM I scored, fan is on the way. Also wanted to catch as many bugs as I can before that as running into big bugs on video looks terrible (but is often helpful to people developing software, or so I've herd).

[Sloth]

Update on my progress:
- Merseine is in.
- AI using Armor Thrull has improved. No longer marked as not playable.
- AI using Orcish Captain was implemented as you've suggested (as removal). No longer marked as not playable.
- Fixed ability cost payment of Night Soil.
- AI using Fungal Bloom has improved (lower priority).

I wasn't able to reproduce the Mindstab Thrull crash.

[friarsol]

Maybe there should be a way to clear out auto-yielded triggers from a menu that way you can interact with the Flotilla when you want to interact with it, and yield when it doesn't matter.

Ah I mispoke here, since I didn't realize it was available. You can clear out your autoyielded triggers from the Forge menu. That should do the trick for you.

[lujo]

Update on my progress:
- Merseine is in.
- AI using Armor Thrull has improved. No longer marked as not playable.
- AI using Orcish Captain was implemented as you've suggested (as removal). No longer marked as not playable.
- Fixed ability cost payment of Night Soil.
- AI using Fungal Bloom has improved (lower priority).

I wasn't able to reproduce the Mindstab Thrull crash.

This is absolutely wondreful! Will test it all out right away! (that is as soon as I figure out how to enable these changes / get the newest snapshot, and then update it with all the stuff I've got in there for testing )

Not sure what you mean by "lower priority", but if it means it'll cast it before it blocks that should be great!

Also, someone or something must've fixed the Mindstab Thrull crash (I've been spamming about it someone must've got fed up with me, or something which was causing it got fixed due to some other issue - it's not crashing the game anymore).

However, if you're up for it, Necrite and Mindstab Thrull are basically the same type of thing, but Necrite isn't allowed for the AI and MST is. However the AI ALWAYS sacrificed either when they go through. The only real problem with them is that MST will get sacrificed when the opposing player has no cards in hand, and Necrite will get sacrificed to kill any creature at all.

Is there a way to limit this behavior or make it more inteligent? In theory MST should be easy - if opponent has less than 2 cards, don't use the ability, there's probably some "discard" tag in there already. As for necrite - a priority target system will be needed for removal at some point anyway, but what tools are currently available to evaluate whether it's worth using removal on something or not?

Maybe there should be a way to clear out auto-yielded triggers from a menu that way you can interact with the Flotilla when you want to interact with it, and yield when it doesn't matter.

Ah I mispoke here, since I didn't realize it was available. You can clear out your autoyielded triggers from the Forge menu. That should do the trick for you.

Thanks, I'll look into it!


----------

Currently working on - a robust and easy to use "Forge Token" system, making further notes on easy to adress stuff, and conisdering making an art set which would improve the drafting experience.

Oracle Wording Art Set | Open

See, the thing with drafting old sets, one of the reasons FE got stuck with a bad rep (same thing happened to Ice Age too), is that the wording on the cards was often clumsy and verbose. Oracle wordings cleared this up, so if you have card art with oracle wording not only does it have the creature types sorted out, it also makes many cards from this era MUCH easier to evaluate at a glance. I don't mind simply making all the FE cards into old-frame oracle-worded arts, if I can just find the vectored expansion symbol with rarity designations for FE. This would also help with spotting these cards as useful and usable when you get them in "full pool" boosters during quest mode where people skip over unfamiliar cards with unreadable textwalls, and I consider it well worth the effort.

Not to mention it makes life much easier while drafting - often people will pass a stupidly worded gem in favor of a meh vanilla creature just because it's easier to tell what the other card does. It's also not too much work for me, I kinda always wanted to do this and feel sorry that the original maker of Forge didn't consider the Shandalar approach to Card display with the dynamic textbox and stuff.

Not to mention that proerly marked empty-silver-gold rarity markings on old sets are a huge boon for drafting and general card eveluation. It doesn't always have to mean much, but being able to tell at a glance that there's a rare missing in a booster you've been passed is important info (to only scratch the surface). I know the "Oracle text box" allready conveys this info, but having it displayed on the more visually appealing box or even obvios without taking your eyes off the booster is great help (and making a video of a draft is much easier when you can just make the booster visible on the screen instead of having to have 2 additional boxes displayed all the time)

-
-
-

Oh, and if Mersiene had it's two missing components coded up in a non-sepecial-case way, apart from applying it to the few things that were exceptions is it possible to implement anything else that needed that code now? Just a though?

[Marek14]

Merseine was "Uncategorized", meaning it needed code that no other card requires, so it probably won't enable anything.
Now, only Raiding Party is missing. This is in a "PayCostEffect" group since the resolution of its ability is pretty weird.

To wit:
First, each player gets a chance to tap any number of white creatures.
Next, each player who tapped some number of creatures gets to choose up to 2*number Plains.
Next, each Plains that wasn't chosen is destroyed.

Even if weak, it would be a fun card to have -- and it might actually help with something else. I guess the problem is that there is a cost paid during the resolution of the ability and the details of how that cost is paid matter. The group contains more cards:
Errant Minion from Ice Age which prevents damage from its ability depending on how much mana was paid.
Kilnspire District from Planechase 2012 which deals damage equal to mana paid (maybe the code for Flameblast Dragon would work?).
Liege of the Hollows from Weatherlight which gives each player tokens equal to mana he paid.
Plague of Vermin from Shadowmoor which combines bidding-like life payments with tokens equal to mana each player paid.
Power Leak from Alpha/Beta/Unlimited/Revised/Fourth Edition which works the same way as Errant Minion
Preferred Selection from Mirage -- not sure why it's here as the details of payment don't really matter for this card, only whether the cost was paid or not.
Worms of the Earth from The Dark which has two different "costs" a player may pay on resolution of the ability.

[lujo]

^ Well, as much as I'd love to see it happen, porting Raiding Party looks intimidating and I'm too happy that we got Merseine to pester anyone for it specifically. I sure would be happy if and enterprising coder did port it in a fit of insanity and thus enabled all those other cards to sneak in too ^^

A small uptade with some stuff I became aware of while making HQ arts for every single card in the set with the Oracle test (and I've also hand-made the expansion symbols with rarty display, allthough I'm not the happiest with the shade of silver and gold in the program I'm using).

- Apparently, the different arts as listed in the game directory aren't ordered the same way in Oracle. Minor thing, but who knows why?
- One thing I seem to have forgotten about is that Combat Medic is a soldier (now a Human Cleric Soldier, apparently), which has a bit of sinergy in the set
- Delif's Cone and Delif's Cube are in fact creations of a dude called Delif. All my life I though it was some woman called Delfi. You misread something once and it totally follows you. Well, a semi-literate buddy misread Wild Mongrel, and he still thinks it's a Wild Mongler. (Not a native english speaker).

And the important one:

#83|Delif's Cube|U|FEM this is the reason I believed Delif's Cube was uncommon, when it's apparently a rare. It's listed correctly in the edition set though, but this is just in case it's affecting anything being listed as uncommon here in the pick list (unless it's a screwup of my very own which isn't out of the question).

Can't wait to show you the new arts, I really tried my best and put in a bit of effort so I got the old frames and textboxes to be wider, got the updated creature types on the cards (With sometimes important info like the Dwarves being Soldiers and thus comboing with Icatian Liuetenant), cleaning up all the refferences to Interupts and generally making the cards less verbose and easier to look at

[friarsol]

#83|Delif's Cube|U|FEM this is the reason I believed Delif's Cube was uncommon, when it's apparently a rare. It's listed correctly in the edition set though, but this is just in case it's affecting anything being listed as uncommon here in the pick list (unless it's a screwup of my very own which isn't out of the question).

I don't believe Fallen Empires had any "official" rares. Delif's Cube is a U1, which is as rare as you can get for these sets, so a lot of times they are referred to as "rare".

FE has 8 card boosters: 2 Uncommons and 6 Commons. Some commons are more common than others (typically with more card art), some uncommons are more rare than others. I know back when FE flooded the market and I was buying $1 packs of it, there never seemed to be a "rare" in the same way that the Alliances packs that I bought did, but at the time I didn't realize how the boosters were distributed.

[lujo]

^ I remember this too, but there were 3 distinct rarities with one corresponding to the current rare, one to uncommon and one to common (that is there were commons with 3 arts and commons with 4). What wasn't the same, I believe, is that you had two uncommons but a lower chance of scoring a "higher rarity" uncommon, which would make it a "rare".

And the oracle makes a distinction which is actually pretty accurate. The expansion list does rank it as rare, while the "pick" list marks it as an uncommon (and does mark other rares as rares, this is the only one discrepancy I've found so far).

Also, while doing this art bussiness I remembered how they messed up and printed all the non-basic lands with red non-basic backgrounds in FE making them all look samey (they also didn't have flashy arts on them). I'm keeping to the old design as much as possible (well pretty much completely besides the creature notation, the tap symbol and all that), but I did fix that thing. It actually looks much better.

EDIT: But here's a thought! If we could confirm without a doubt that the original set really had no rares and that all uncommons were equaly legit, we could play around with the rarities to make the draft way better. However, I remember opening 2 boxes of FE once and getting a load of stuff that's common and uncommon on oracle but not getting more than 1-2 of the rares. So I'm kinda certain there were proper rares in the set, even if they were called "uncommon X".

[Sloth]

EDIT: But here's a thought! If we could confirm without a doubt that the original set really had no rares and that all uncommons were equaly legit, we could play around with the rarities to make the draft way better. However, I remember opening 2 boxes of FE once and getting a load of stuff that's common and uncommon on oracle but not getting more than 1-2 of the rares. So I'm kinda certain there were proper rares in the set, even if they were called "uncommon X".

The uncommons of FE were printed with sheets of 121 cards (121 is still the normal size of sheets btw). 25 cards appeared 3 times on the sheet, 5 cards appeared 2 times and 36 cards were present only once (the rares).
Delif's Cube is a U1 (it appeared only once on the sheet), so it should be considered a rare.

Source: http://crystalkeep.com/magic/products/fallenempires.php

[lujo]

^ Yep, that corellates with my experience too. And also makes sense. It's a bit of a shame, but it makes sense.

Allright, so, here's a visual spoiler of the first draft of a FE graphics/wording/templating update. I think I can do better than this so feel free to pick it apart with comments, dismissals, suggestions and info on how to make it more in line with Forge for (possible) integration. I would absolutely LOVE to have this polished and become standard issue for forge - it may not look like much but it helps a TON in making the older sets more accessible beyond the obvious legendary cards everyone knows about.

http://www.filedropper.com/pics_9

I took liberties with the oracle wording of 2 cards only, Homarid and Tidal Infulence, as the Oracle wording made those two even more verbose and confusing, so I used the new Tarkir bullet list template to hopefully make them easier to understand.

I can rearange all the elements, if I'm off a pixel or two, make sure you tell me, and the copyright line needs to be made exactly right. I've also contacted the guys who made the program to see if the uncommon and rare expansion set symbol fillings can be made darker (I just made the expansion symbol the program automatically filled it up and it's too bright I think). Also, I can only do so much about the arts - some are grainy because they were made that way, but I can go through all of them again and see what I can do to make them less grainy. It's a shame the spoiler starts with white as it has the lousiest art in the set, while some other colors and cards benefit quite a bit from this kind of treatment.

I'd personally use this template for all the old cards, it's more spaceous, the rarities are obvious, the references to interrupts and unnecessary lines of text on old cards just confuse people, you can get the creature types properly updated for better visibility and easier integration into your quest decks etc etc.

Would be a ton of work, but if I could get the template to work, I wouldn't even mind doing it. Provided I would be able to actually make all the set symbols, the program is very iffy with those (I dread attempting to do Ice Age and Alliances expansion symbols).

-----

And - is there a new snapshot with all the developments? If I got that, and adjusted everything and added the pics, I could do a video of a test pod and finally show you guys stuff instead of just blogging here XD

EDIT: I ajdusted the text boxes a bit because text was crowding on the bottom lines; I corrected the card image area to include the 2 missing lines of pixels on the left, lol, now pictures fit snugly into the box; I moved the card type text up 2 pixels, it was too low; gonna try to smooth the artworks out and try to see what controlls the silver and gold coloring on the symbols to darken it a bit, then I'll put up an improved visual spoiler.

[Sloth]

Update on the AI:
- AI using Spore Flower has been improved. The AI will use the ability instead of chump blocking.
- AI using Necrite has improved (it will only destroy better creatures). No longer marked as not playable.
- AI using Mindstab Thrull has improved (it will only sac him for 2 or 3 cards).
- AI playing with and against Icatian Scout has improved (it will anticipate a creature gaining First Strike).

[lujo]

^ Awesome! The thrull bussiness is great as they're common and being thrulls are important fodder (and a well played Mindstab Thrull is scary s**t, getting hit for 3 cards after a Hymn to Tourach hurts like all hell, and getting beaten up by subsequent ones is perfect AI play). The Icatian scout thing is deceivingly important - with so much defined by creatures having that thing interpreted right by the AI is actually huge (as crappy as the card looks otherwise). Great Work!

art status update | Open

I've adapted the template I'm using for art to be more in line with the one used by Forge so scrolling between cards using "original scans" and it isn't jarring, and I think I can make templates with less eye stabbing rarity symbols for each rarity. Otherwise I'll just handstamp a better looking colored expansion symbol on each art XD Also I'll try to make the white lettering a bit more gray to stand out less and be less crisp (the FE - Alliances cards are darker and more "washed out" in general, so really crisp colors on them don't look so good). I've also compared the art I've done with the one from HQcardgenerator - and while my art looks kinda terribly grainy, the HQCardmaker art looks awfully blurry. The reason is it automatically decreases sharpness on every crop and that doesn't actually work right with a lot of these old-school arts - I've just overdone the opposite (sharpening) on quite a few.

---------

Here's some more notes on AI stuff that could use looking into:

Simple (?):

- The AI doesn't seem to activate the important first strike ability on Order of Leitbur and Order of the Ebon Hand (and doesn't seem to realize it has the capability).

| Open

I've had situations where I attack with an Orgg into 3 "pump" knights, and the AI chump blocks with 2 (Goblin War Drums), loses them both and I only realized after the fact that it had 3 ways to kill the Orgg with no losses (block w all 3 and add first strike, or block with 2, pump and add first strike.

I realize there were issues with AI spamming pump and activated abilities on everything, but first strike activated abilities truly only need to be activated once. This goes for Icatian Infantry, too.

- Thrull Wizard is playable, except the AI doesn't seem to be using his ability.

| Open

If there's a simmilar AI packer for cards like Deathgrip, Lifeforce, Douse and such, maybe it can use that? If there isn't, maybe there could be so it can be made available for all those too?

- Ebon Praetor looks complicated, but the only thing the AI isn't doing is removing counters.

| Open

This is a bit tricky though, as you need to first resolve the "add the -2/-2" counter and then the sacrifice or you'll end up not removing anything, down a creature and stuck with a 3/3 Ebon Preator. Otherwise if Lord of the Pit is operational, it's basically that but with a taste for thrulls (tokens a priority, then uncommons (basically only Thrull Wizard in the set), then Basal Thrull, then all else). Lord of Pit AI ought to make it playable on it's own, though.
-
- Ebon Praetor on an empty board is a bad move by the AI though, I've got note that sayes that.

- Homarid Spawning Bed also looks tricky but is actually really straightforward - if a blue creature's about to die and isn't a token, turn it into tokens. If the AI would do that it's all there is to it

benefits | Open

(and there can easily be a really strong constructed deck in quest mode with Viscerid Deepwalker / Errant Ephemeron / AEthersnipe / Peregrine Drake / Mulldrifter + a controll spells and it. The reason its not more well known is that blue only became a creature color much much later, but with stuff you can cheat into play for less than it's casting cost and/or flash that thing is complete cancer. In a right deck you can get 6 tokens for U out of Somber Hoverguard o.O That is if someone tries to kill your 3/2 flyers you got for U. That's not even going into Tidal Visionary / Painter's Servant + evoke (say, Ingot Chewer or Shriekmaw + Homarid Spawning Bed nonsense or straight up faeries / AEther Vial). It's one of the silliest cards ever printed in MtG, tbh, so simple and straightforward and so utterly broken)
-
- A potential issue with that approach is the same that Armor Thrull might prove to have, and that is if you block with the creature and the AI chooses to sac it instead of trading (once you select the order in which damage is dealt, this has happened with Armor Thrull and seems to be the only minor bit of issue with it. It's less of an issus with HSB because you generally end up with more power on the board so you can trade for some of the tokens next turn).

- Balm of Restoration seems playable, but the AI tends to use it to heal itself if it's below 20. This is a very suboptimal use, the combat trick aspect is what's it all about. A tweak to when it's willing to use it (below six or so to take it out of Goblin Grenade range) and it's probably fully playable.

- Zelyon Sword and Spirit Shield, the old "equipment" - the AI isn't activating them. Drdev (or was it friarsol?) mentioned something about recoding them to work more like equipment which they essentially are. This would make them usable by the AI and that's good as they're both high picks.

bonus benefits | Open

If there's workable AI for these things it'll also work on Ashnod's Battlegear, Tawnos' Weaponary and all the other old "equipment"

- The AI will use the Icatian Javelineers ability on the opponents life total. Bad move, should only be done if it's about to die and it can't kill anything. That thing is otherwise format-defining in many ways. If it never ever shot "to the face" but did save the counter to kill opposing creatures it would still be format defining.

- Goblin Chirurgeon's ability isn't being used, if the AI was willing to use it to save combat creature's he's playable too.

- BUG? Farrel's Mantle and (possibly) Farrel's Zealot damage can't be prevented for some reason (or so it seems).

Tricky (?):

- On a rather tricky note - Goblin War Drums seem to be confusing the AI. I think the "Has to block with 2 creatures" (present on other cards in the game, but Goblin War Drums is a staple in the set) confuses it's chump-blocking routine. What often happens is that it turns a situation where there's no difference between chumping and proper gang-blocking, but the AI will do it's best to force a chump block (which under the circumstances looks ridiculous and plays out catastrophically). Keep in mind that disabling Gobling War Drums for the AI is out of the question, that thing is impossibly important in the environment.

- Question: What exactly determines deckbuilding choices? Allowing more cards and fixing the list did wonders, but the biggest problem now is insane splashing choices (Feral Thallid makes a really bad "only green card", especially with only 1-2 forests in the deck if any at all).

- The AI will frequently go for 4 colors without appropriate lands, and the "sac" lands seem to be perfectly playable (might need some tweaking to make them useful, like the AI using them to rush cards with 5+ cc). It also tends to stick them into any deck, as it doesn't seem to realize their "color identity" at all.

- The "limit X" thing to make the AI stop drafting/spamming particular utility cards while deckbuilding in the same vein as remAI also has to be a thing before I can really make it all work (you get 5 Goblin War Drums or 5 Elven Fortress in decks).

EDIT: - The AI tends to go for 10 or so non-creatures unless going white which has more creatures. This in combination with spamming redundant utility is causing it to misevaluate it's strong colors (in some cases quite badly). The general rule is no more than 6 or 7 non-creatures in a draft deck under most circumstances, so some way to experiment with this would be highly appreciated.

- There REALLY needs to be a splashable tag simmilar to the "remAI" tag. Trying to get any set to draft right without it is kinda impossible.

EDIT: Also, if all the "simple" ones get done, and some progress is made on the "tricky" ones and the ones already done turn out to work fine - this thing is totally going places!

some thoughts from Mark Rosewater of all people | Open

EDIT: Also, lol at this tweet from Rosewater: https://twitter.com/maro254/status/312656329623875584

And this one is also illuminating - https://twitter.com/maro254/status/312620400511156224. V important set Fallen Empires, it's actually a terrible shame how many factors contributed to it developing a bad rep. This one makes me interested in possibly attempting to "fix" issues with rarities in it. Not high on my priority list, but if this project gets done (and it's going great so far) I have half a mind of buzzing Maro to talk about FE design a bit. Seeing how much love for it was all over Time Spiral you can tell this guy had a thing for it. (Wouldn't want to draw his attention to Forge, though, so I won't do it, but I'm 100 percent sure that if we got it to draft right we'd make him fall out of his chair XD Not to mention he'd be spamming Forge FE drafts in private, there's no way the guy who made Time Spiral wouldn't XD)

[lujo]

Right so here's some art after a bunch of work on it. I can actually fiddle with expansion symbols and rarities better as I've delved into the .txt settings. Whish I knew the first thing abotu graphics, but I don't, so this is the best for now. Needs more tweaks, but if you've got any suggestions, they're all wellcome (and, err, I'd laugh if it turned out this is the default setting for Forge if you don't download the actual arts). Also whish this macihne I'm on could run Photoshop or Gimp so I wasn't doing all this in paint -.-'

Some Art! | Open

-
- Maybe black borders on Expansion Symbols for uncommon and rare? White cards are a problem as usual.
- Maybe just even darker uncommon and rare symbols? The oracle rare symbols for FE are almost orange...
- Even more gray text? (Can't figure out how to make it not be mono-color but filled)
- I do need to adjust the location of the p/t to be in line with forge
- Even these templates are a few pixels off from what forge art uses (at the top and the bottom). I'll have to shave those and readjust it seems.

- Is my Homarid rewording helpful even if it breaks sentence structure a tiny bit with the last line (maybe)?

[sweetLu]

Tree tide counters??? How many is that (the wording looks good besides the tpyo)

[lujo]

Tree tide counters??? How many is that (the wording looks good besides the tpyo)

XD Fixed all 4 Homarids, Ty!

Also added a bunch of rares in different colors up there in the spoiler so you guys can check them out and critique / give advice.

And here, try this for typo's and wording conundrums:



Actually even if this wording is problematic (you can't really phrase it right in English) it's still way more readable than the original one, and the card is one of the biggest "TLDR I won't even bother" cards in the history of magic. Yet, with a Homarid Spawning Bed out, or if you've been stacking Vodalian Soldiers behind a Vodalian War Machine, it kills the other guy dead.

(Also, I hope noone gets mad because I'm sharing images this way and stuff, not sure if this is the best way )