1st Official Wagic Deck Competition - TALK

[flad_nag]

thank you very much!

[whismer]

Is it possible to see what happen in the tournament?

Ex : A tree with the confrontation 12 vs 1 11 vs 2 or Round Robin Group and Result

It's wil be funnier to see the global competition that only result

What's the formati did you use ?? 3 of 5, Double-Lost Elimination?

Thx

Whismer

[Psyringe]

Is it possible to see what happen in the tournament?

Ex : A tree with the confrontation 12 vs 1 11 vs 2 or Round Robin Group and Result

Yes, I'll give the results of the individual matchups. The tournament will also stretch over several days, so people can follow and discuss the performance of the decks.

I haven't decided on the fixture yet, it depends a bit on how many submissions I get. I'll probably do a group stage first and a knockout stage afterwards (like in a soccer world cup ) - this has the advantage that every deck can present itself in a guaranteed number of matches (in the group stage), but we also get the suspense of the single elimination stages up to a real final.

For a statistical analysis, it would be better to just put all decks into a huge round-robin group, as I do for my AI comparison statistics. But for a tournament, a combined group stage / knockout fixture will be more fun, imho.

[Psyringe]

Update for all people using my "AI Test" mod: One of yesterday's SVN updates required an update to the mod as well. The updated files are here.

Status update for the competition: We're now at 15 participating decks - keep them coming.

I also decided to admit a deck that was already posted on this forum - it was submitted by the same person who posted it two weeks before, in a largely overlooked thread, so I figured that this deck can still count as "new for Wagic". That's a one-off decision though, also based on the reasoning that two weeks ago he couldn't have known that this competition would take place.

[Psyringe]

Competition status: So far, 13 participants submitted 17 decks in total. I've checked all decks now and suggested adjustments where necessary (some decks were not Vintage legal or used cards not supported by Wagic). By now, everyone who submitted a deck should have received a reply from me. If you haven't, please check your forum inbox. If nothing's there, then your deck hasn't reached me, in that case please submit it again.

Keep the decks coming - submissions can be sent until Saturday, 23:00 GMT. I think we'll have a very interesting tournament, don't miss your chance to become a part of it.

[Niegen]

Errr...

Reading Psyringe's last comment, I want to tell candidates that building a working AI deck currently is HARD. You MUST test it, and more than once.

The AI is really dumb. If you don't test it, it WILL underperform. Don't take for granted that adding a good card to the deck will make it more powerful.

The current Elfball deck for example went through several stages before reaching its current state. It is tuned. Some card choices are probably not optimal from a human point of view, however, the deck is built as "foolproof" as possible. No matter what the AI does (barring playing Elvish Promenade with an empty board, which has a low chance of happening, but can happen nevertheless), there is no "bad play".

It does not even play removal, so it cannot "goof" and remove an irrelevant target, wasting a card and potentially a turn. (Green does not have good "creature removal" anyway).

I am currently working on 6 or 7 decks, but expect to only submit 2, or maybe 3 of them, due to lack of time and testing. The AI does not play the others as it should.

[Psyringe]

Sound advice.

Nevertheless I think that there's room in this competition for "experimental" decks, or even "let's just throw this together and see if it works"-type of decks, as well as the well-tested, solid high-performance decks. The experimental decks aren't doing any harm, and this competition is meant to be fun anyways. So if people want to participate by just submitting the deck that they like to play themselves and see how it fares, more power to them. It probably won't win the tournament, but it will add to the fun.

Also, keep in mind that the AI is constantly developing. For example, the recent fix to AI memory does make the AI better at removal. Is it now good enough with it to warrant the inclusion of removal cards in high-performance AI decks? I have no idea (yet). But playing decks with and without removal against each other will give me a hint. That's why I'm very curious about the submitted decks, and that's why I'm encouraging people to stray away from the well-trodden paths, and/or submit decks even if they aren't sure how the AI will play them.

But of course, if someone wants to *win* the competition (or at least get into the quarter finals), thorough testing is definitely advised.

[jeremy]

there is the option of building a deck with just creatures,spells that make creatures, lands and fetch lands.
that way as long as the AI plays the creature, you can fill it with coming into play abilities.
this was the way I approached one of the decks I sent in with the ZEN allies theme of something happening when an ally hit the deck, the only "spell" that needed targeting was swords to plowshares, which was cool since there was potential for massive life gain anyways and the added life you gave shouldn't matter.
This was the mono white version of it, I'll work on a green/red version once I free up some time this week, maybe even splash in blue since with fetch lands, you get around the AI inability to use dual lands and you thin out the deck when used too. All that and it hits landfall if anything nasty is out as well. Perhaps I should add Emeria Angel hmmm....

[Psyringe]

Update: With 50 hours to go until the competition starts, deck count is now at 24.

I know that at least one more deck is currently in testing, so we'll have 25 decks minimum. If I don't receive any more decks, then the fixture for the competition will look like this:

First stage, 5 groups of 5 decks each, first 3 of each group get promoted to next stage
Second stage: 3 groups of 5 decks each, best 3 of each group (except the third-placed team with the least points) get promoted
Afterwards: Quarterfinals, semifinals, final as one-on-one knockout stages.

If I receive more decks (but less than 7), then we'll have a competition of 32 decks. I'll add some experimental or themed decks of my own to get the count to 32 if necessary - the decks I add won't be high-performance decks, but shouldn't be pushovers either, in any case it's nicer to have a full field of competitors than giving some decks a bye. With 32 participants, the tournament will go through these stages:

First stage: 8 groups of 4 decks each, best 2 of each group get promoted
Second stage: 4 groups of 4 decks each, best 2 of each group get promoted
Afterwards: Quarterfinals, semifinals, final as one-on-one knockout stages.

Matches will be played as "Best of 7" from the quarterfinals onwards and "fixed N games" before, I'm currently contemplating on the exact value for N (I want draws to be a possibility in the group stage). Before the competition starts, a large number of fast-paced "training matches" between random competition decks will be played so that the AI can get accustomed to the decks.

[AzureKnight]

My only complaint is that the AI in Wagic doesn't really know how to play the decks properly.

And just can't use certain cards altogether.

For example, the pain lands, filter lands and fetch lands all can't be used.

This seriously restricts deckbuilding creativity.

Along with the inability to mulligan, using a largely basic manabase is bound to lead to mana screw in multicolor decks.


Still, I think this idea is actually really great.
We can seriously improve the types of AI decks available in Wagic by holding such a competition regularly.

Keep it up!


(Sending my list soon, hope it's not too late!)