Artifact

pew pew pew
Post Reply
Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Artifact

Post by Jonathan »

It's a thing? Or is it? I can't find out if it's playable by plebes such as myself.

Edit
FAQ
http://drawtwo.gg/

Alan
Veteran Doodler
Posts: 2758
Joined: Fri Jul 18, 2003 2:32 am
Location: Where I am
Contact:

Re: Artifact

Post by Alan »

It’s not out yet. I am considering trying it, but may just stick with TESL. There’s no free to play option with Artifact, you have to buy packs.
Image

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

https://www.pcgamesn.com/artifact/artif ... ling-cards

Dev alludes to the announcement of an open beta before launch Nov 28, but no date yet. The beta is running right now, but it is a closed beta.

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

Just exploring hypothetical mana curves. Even with lots of small cards, you often wind up with lanes where you can't play a new card. Developing improvements, using abilities, and buying items is extremely important.

Deck Format: 1s 2s 3s 4s 5s 6s 7+s

Game ends in Turn 8 (max 155 mana)
The final optimal deck:00 00 00 03 14 08 15
The expected mana spent for this deck:114.95207(min: 89)

Game ends in Turn 7 (max 126 mana)
The final optimal deck:00 00 00 09 12 08 11
The expected mana spent for this deck:95.29912(min: 75)

Game ends in Turn 6 (max 99 mana)
The final optimal deck:00 00 00 17 08 08 07
The expected mana spent for this deck:76.73265(min: 59)

Game ends in Turn 5 (max 75 mana)
The final optimal deck:00 00 04 15 10 07 04
The expected mana spent for this deck:59.06466(min: 42)

Game ends in Turn 4 (max 54 mana)
The final optimal deck:00 00 12 14 09 05 00
The expected mana spent for this deck:43.347255(min: 32)

Game ends in Turn 3 (max 36 mana)
The final optimal deck:00 00 21 12 07 00 00
The expected mana spent for this deck:30.48043(min: 18)

Alan
Veteran Doodler
Posts: 2758
Joined: Fri Jul 18, 2003 2:32 am
Location: Where I am
Contact:

Re: Artifact

Post by Alan »

How are you calculating those mana curves? Why do they all give decks that have zero 1 and 2 mana cards?
Image

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

I ripped the algorithm from some dude for MtG, then I modified it to handle the Artifact ruleset. The algorithm is doing an exhaustive search over deck combinations, but uses Monte Carlo simulation (n=200000) to calculate the average Mana spent since a closed form is too hard to write. The exhaustive search is too time consuming, so he eliminates some bad deck combinations using a gridded search. It reruns the exhaustive search near the outcome of the gridded search until it converges on a result.

The code is still pretty slow and shitty Java. I might just rewrite it in C++ so it can do 8s and 9s. I think it would be interesting to have top N decks instead of top 1, too.

I'm assuming you play your biggest card(s) in hand in each lane, and then i go down the list in order. I'm giving the odds of an extra draw as 1/6 every round because I think that's the odds you get offered a Potion of Knowledge in the shop. After all the lanes are exhausted, you draw two cards from your deck and repeat. Games are a fixed length of turns.

I'm ignoring color, disruption, and hero death.

The first turn of Artifact, you have 3 Mana per Lane and 5 cards in hand. So, 1s and 2s don't help you stay on curve because you already have 3 Mana. Unless you have significant card draw in your deck, you quickly exhaust your hand if your deck has 1s and 2s and fall behind on Mana spent.

In actual play, 1s and 2s seem good for triggering procs of passive hero abilities, but significantly weaker than 3s and 4s. It's common to build a deck with zero 1s and 2s. It's not uncommon to spend zero Mana on the first turn, either. You really need a lot of 3s to get anywhere near 9 Mana spent your first turn.

Because heroes, melee creeps, and items enter play without use of Mana, you can develop board state without playing on curve in this game.

Alan
Veteran Doodler
Posts: 2758
Joined: Fri Jul 18, 2003 2:32 am
Location: Where I am
Contact:

Re: Artifact

Post by Alan »

Didn’t realize that the games start on 3 mana per lane. That definitely makes 1-2 mana cards a lot less valuable and probably only worth including to take advantage of certain synergies.
Image

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

Beta only ten days before launch on Nov 28.

https://www.artifactfire.com/artifact/n ... discovered

Alan
Veteran Doodler
Posts: 2758
Joined: Fri Jul 18, 2003 2:32 am
Location: Where I am
Contact:

Re: Artifact

Post by Alan »

I've been reading articles (drawtwo.gg has some really good stuff) and watching Twitch/YouTube. I ended up pre-ordering even though you don't get anything extra for doing that.

This game looks really interesting. But it looks like it has a very steep learning curve and very very high skill cap.
Image

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

I have also pre-ordered. I agree with you!

I think keeper draft isn't the play for building your collection. I think you pop the free packs, go to the marketplace for your initial deck list missing cards, and grind packs through expert constructed gauntlets.

The link seems good. I also like Swim's stream and card ratings. I put some reference links in the first post.

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

Here's what I've been working on for a day 1 purchase. I'm curious how much this will cost.

http://playartifact.com/d/ADCJTIL6bkCBU ... dGVnZXJpZQ__

https://www.artifactfire.com/artifact/d ... gerie-2232

My observations (not from play obviously, nor guides):
"Tempo" exists in this game, but slightly differently than other games. Some cards, once played, demand your opponent respond. You can play these cards even if you need initiative on another lane, because you'll expect the response. Most cards don't have this quality, and some only have it situationally (like direct damage to bring a hero into the danger zone for Thundergod's Wrath).
Lane position is random initially but deterministic. Manipulating lane position is super important. I see lots of streamers clear their opponents' threats without actually getting the damage in to the tower, which is how you lose games.
Constructed is supposedly a fairly fast format. I have seen a few turn 3 wins but the usual sort of fast game is over by turn 5. Turn 6 or 7 wins are very common.

With that in mind, I put together a deck that is designed to get it in without going wide. It has lots of tools for poking holes into a board and dealing damage in there. It's optimized for a quick win, but PA and Tinker give a few control options if it doesn't come together immediately.

Probably I'll change lots of stuff, even the heroes, once I get the game, but I think I'll probably iterate on that idea for a while.

Alan
Veteran Doodler
Posts: 2758
Joined: Fri Jul 18, 2003 2:32 am
Location: Where I am
Contact:

Re: Artifact

Post by Alan »

I like this explanation of the importance of the various resources at different stage of the game.

I'm wondering if certain cards, particularly Drow Ranger, will be overpriced for the first few weeks after release. It seems like she has been hyped a lot as one of if not the very strongest hero in the game.

I'll probably play around with whatever cards I start out with until I get a feel for the game, then try the Draft modes for a while.
Image

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

jonathan@nofucksgiven:~/Downloads$ tail -n4 turn3
The final optimal deck:
0 0 20 13 7 0 0 0 0
The expected mana spent for this deck: 30.4707(min: 15)
----------------
jonathan@nofucksgiven:~/Downloads$ tail -n4 turn4
The final optimal deck:
0 0 12 13 10 5 0 0 0
The expected mana spent for this deck: 43.3485(min: 30)
----------------
jonathan@nofucksgiven:~/Downloads$ tail -n4 turn5
The final optimal deck:
0 0 5 14 10 7 4 0 0
The expected mana spent for this deck: 59.1566(min: 42)
----------------
jonathan@nofucksgiven:~/Downloads$ tail -n4 turn6
The final optimal deck:
0 0 1 14 10 7 6 2 0
The expected mana spent for this deck: 77.2361(min: 60)
----------------
jonathan@nofucksgiven:~/Downloads$ tail -n4 turn7
The final optimal deck:
0 0 0 8 12 7 6 5 2
The expected mana spent for this deck: 97.8896(min: 74)
----------------
jonathan@nofucksgiven:~/Downloads$ tail -n4 turn8
The final optimal deck:
0 0 0 4 12 8 6 5 5
The expected mana spent for this deck: 120.726(min: 97)
----------------
jonathan@nofucksgiven:~/Downloads$ tail -n4 turn9
The final optimal deck:
0 0 0 8 12 7 6 4 3
The expected mana spent for this deck: 157.838(min: 100)
----------------

I redid the optimal mana curves with 8 and 9 drops. I get a off-by-one difference comparing the C++ version with the Java version. Both have essentially the same mana spent expectation (you can change all these curves by a couple cards and still get essentially the same mana spent on average) so I'm chalking it up to differences in RNG. I think I need to examine variance, look into curves forcing 3 1s (J'Muy) or 3 2s (Legion Commander), and see if I can't consider the difference between mono and multi color. Tri color is pretty common in draft, and I suspect it means you need to have more lower cost drops.

Interesting observation is you always want about 10-12 5 drops unless you're a very aggressive deck.
Weird fact: no 6 drops in either black or green (except Curse of Atropy and Ravenhook, which are both super situational)?

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

http://playartifact.com/d/ADCJSIZJLkCCh ... WNlU21hc2g_

I like this beta player's BR Aggro list.

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

Ugh, apparently you need to authorize your credit card for the community market separately from, you know, buying shit on Steam.

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

I spent $28.37 on my deck, including $9.75 for Drow Ranger. Second most expensive card was Tinker at $4.
I've seen Drow for as low as $7.76 and as high as $11.
Axe is running about $12 (and going up).
The in-demand creeps like Emissary and Avatar are a couple bucks.

Alan
Veteran Doodler
Posts: 2758
Joined: Fri Jul 18, 2003 2:32 am
Location: Where I am
Contact:

Re: Artifact

Post by Alan »

Looks like the first few copies of Axe sold for nearly $40, and the first few copies of Drow sold for around $25.

$25 for a meta deck is not bad. I'll have to take a look tonight.
Image

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »


Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

An item you listed in the Community Market has been sold to Даша ^.^. Your Steam Wallet has been credited 0.03 USD.

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Re: Artifact

Post by Jonathan »

I guess there's not going to be a Chinese meta and a EU meta and an NA meta in this game. Just today, I've interacted with dudes/dudettes from:
  • Canada
  • China
  • Russia
  • I feel like he/she was likely Korean but the name was Engrish

Post Reply