Manalink Tournament (May) (tournament complete)

[Stroggoii]

Awesome tournament, I must say I'm really happy and kinda impressed about the deck's performance, it's pretty obviously not a finely tuned deck, but I guess Kor Firewalker did become an MVP in a red-heavy meta.

About what I'd like as a price, I really, really want Sarkhan the Mad. This card is awesome and I'm currently building Bloodquake around him in cardboard, it'd be great to have him in manalink.

[aww1979]

A bunch of statistics:

Number of decks submitted: 14
Number of players participating: 11 (kakodrilo, enigma73 and thedrigo did not test decks)

Decks Most Often Seen Played By AI: (this statistic is pure chance)
1) 62 - stroggoi
2) 60 - enigma73 & shantak (tie)
3) 59 - yggdrasil
4) 58 - thedrigo
5) 57 - brandon
6) 56 - cognis
AVERAGE: 52.0 matches played per player
7) 51 - jatill
50 - kakodrilo
9) 45 - scherbchen & aww1979 (tie)
10) 44 - nyktorion & juzamjedi (tie)
11) 37 - salbei
TOTAL: 728 matches played
Salbei clearly got the short end of this stick. He still managed to tie for the win though, because the AI did very well with his deck.

Most Human Player Wins/win percentage: (I could've ranked this according to win percentage instead of # of wins; both are relevant numbers. Swap 5 with 6 and 8 with 9 for the other way)
1) 80/89=89.9% - aww1979
2) 74/83=89.2% - cognis
3) 73/83=88.0% - juzamjedi
4) 66/77=85.7% - nyktorion
5) 60/73=82.2% - jatill
AVERAGE: 57.5 wins, 82.8% win percentage
6) 54/65=83.1% - salbei & yggdrasil (tie)
7) 54/66=81.8% - shantak
44/57=77.2% - brandon
9) 38/48=79.2% - stroggoi
10) 35/49=71.4% - scherbchen
Players high on this rank are good at borrowing decks they have little experience with and doing well. Alternately, it could mean that they got very easy opponents in the gauntlets, but that would take a lot of work to calculate. A low score here could mean you are weak in one major archetype or prefer to play one over another (aggro or control; we had no combo so that one doesn't apply here)

Most 10-0 Gauntlets (playing with any deck):
1) 5 - aww1979, cognis
2) 4 - juzamjedi, nyktorion
3) 3 - salbei, yggdrasil
AVERAGE: 2.5
4) 2 - shantak
5) 1 - brandon, jatill
6) 0 - scherbchen, stroggoi
TOTAL: 28
(same idea as the previous statistic)

Most 10-0 Gauntlets (all players with specific deck):
5 - juzamjedi
4 - salbei, nyktorion, yggdrasil
3 - scherbchen, jatill
AVERAGE: 2.5
2 - stroggoi, thedrigo
1 - cognis
0 - enigma73, kakodrilo, aww1979, snantak, brandon
TOTAL: 28
A high score here means your deck is easy to learn and that it is a good deck. It also means your deck is fairly consistent.

Points Gained by Gauntlet:
70 - yggdrasil
68 - juzamjedi
65 - salbei
61 - nyktorion
57 - shantak
54 - scherbchen
49 - stroggoi, jatill
AVERAGE: 45.1 points
41 - enigma73
38 - thedrigo
27 - kakodrilo
18 - cognis, brandon
17 - aww1979
Good, easy to learn decks score high here. Poor decks and/or difficult to learn decks score poorly.

Points Gained by AI wins:
85 - salbei
80 - stroggoi
75 - jatill
70 - yggdrasil
55 - nyktorion
50 - scherbchen, juzamjedi
45 - aww1979
AVERAGE: 44.6 points
40 - enigma73
30 - thedrigo
20 - cognis
10 - kakodrilo, shantak
5 - brandon
A high score here means the AI is good with your deck, and that humans are weaker playing against that deck, possibly due to combat tricks, metagaming, and things like that.

How far people got in gauntlet with their own deck:
10 - nyktorion, salbei
8 - yggdrasil
7 - shantak
5 - scherbchen, juzamjedi, stroggoi
AVERAGE: 4.9 rounds
2 - aww1979, brandon
0 - jatill, cognis
People at the bottom of this list are clearly honest :p (Players at the top probably are, too But we cannot doubt jatill's honesty at reporting a zero!)
edit: Corrected Cognis's result in this category; I read off the wrong line, and corrected the average as well. (apparently cognis is honest too )

Number of times players lost to their own deck:
4 - stroggoi
3 - salbei
AVERAGE: 1.2 losses to their own deck
1 - aww1979, scherbchen, jatill, juzamjedi, nyktorion, shantak
0 - brandon, cognis, yggdrasil
A high score here means your own deck is easy for the AI to play and/or surprised you!

Most AI wins / AI win percentage (sorted by win percentage because that's less random than number of wins, which is partially dependent on # of times the AI got to play the deck)
17/37 = 45.6% salbei
15/51 = 29.4% jatill
16/62 = 25.8% stroggoi
11/44 = 25.0% nyktorion
14/59 = 23.7% yggdrasil
10/44 = 22.7% scherbchen
10/45 = 22.2% juzamjedi
9/45 = 20.0% aww1979
AVERAGE: 16.9% AI win percentage, 11.2 wins
8/60 = 13.3% enigma73
6/58 = 10.3% thedrigo
4/56 = 7.1% cognis
2/50 = 4.0% kakodrilo
2/60 = 3.3% shantak
1/57 = 1.8% brandon
TOTAL: 123 AI wins (out of 728 matches)
This is probably the best indicator of how good the AI is with a particular deck. Clearly, the AI cannot play burn decks at all, as the bottom three scores are the burn decks. The two control decks scored highest, which surprised me a great deal. They were designed heavily with the metagame in mind, contributing to their success.

Highest Individual Deck Score: (just a cool statistic :p)
30 points - Yggdrasil, playing with Nyktorion's deck



Percentage of Points Gained by AI Wins:
72.6% - Aww1979
62.0% - Stroggoii
60.4% - Jatill
53.6% - Salbei
52.6% - Cognis
50.0% - Yggdrasil
49.4% - Enigma73
48.1% - Scherbchen
47.4% - Nyktorion
44.1% - Thedrigo
42.4% - Juzamjedi
AVERAGE = 36.6%
27.0% - Kakodrilo
21.8% - Brandon
14.9% - Shantak

Players at the top of this list designed decks more with the AI in mind. Players at the bottom of this list designed decks that the AI is incompetent with. This statistic shows that the AI sucks with burn decks, and they really skewed the average. It isn't necessarily an indicator of the quality of the deck. (my own is at the top yet I only got 10th place) All the prizewinners had at least 50% here, indicating that if you want prizes, focus on how the AI uses your deck.

[aww1979]

Personal comments:

enigma73: I loved the idea of this deck. (Heck, one of the ones I built was based on the same idea!)
stroggoi: Kor Firewalker was one heck of a metagame choice, as was Blood Knight. Blood Moon or Magus of the Moon would have make a great addition here.
Salbei: This deck was incredibly boring to play with :p It obviously did very well, though, both in human hands and in AI hands. On the human side, I'd love to have seen Mana Drain in this deck, but the AI usually plays creatures postcombat, and usually Mana Drain gives mana precombat, so it would go to waste.
nyktorian: This deck gets my vote as the most fun deck of the tournament, because it probably took the most skill to play, as there were always more options to consider in play than other decks had.
scherbchen: This was a fun deck to play, too. Grove of the Burnwillows and Punishing Fire would have also gone with the theme, though last I checked, Punishing Fire's recursion ability doesn't work.
jatill: Also a boring deck to play with :p
Yggdrasil: This was a fun deck. Black is my favourite colour
Kakodrilo: There were some fun cards in here, but some of them didn't work very well together, like Price of Progress and Magus of the Moon, which was a bit frustrating at times.
Cognis: I like goblins! The idea of adding black was neat, though I lost several games because of the basic swamps and not having enough red mana in opening hand. More black/red duals would have helped a lot here, though keeping 2 swamps is important due to the Blood Moon in here.
JuzamJedi: I liked this deck, too. I'd personally have played Ajani over Elspeth (I designed a very similar deck doing that, but I didn't include it due to your deck being so close)
thedrigo: The addition of green to white weenie made this deck unique. I absolutely *hated* the Mox Diamond in here though :p
aww1979: My goal with this deck was to catch a few players off-balance with the discard. I was hoping it would end a few gauntlet runs with quick, massive discard. It did do that, but humans didn't win much with it; in fact, it had less human gauntlet wins than any other deck.
shantak: I really liked the idea of this deck; I'd never thought of using Oath of Druids with those one-turn creatures before. Unfortunately, the AI couldn't grasp the basic concept :p
Brandon: Ugh, sorry about last place If I had one suggestion to improve, it would be to include more versatile cards. For instance, choosing Burst Lightning over Lava Spike, since it also targets creatures and can be kicked.

[aww1979]

About what I'd like as a price, I really, really want Sarkhan the Mad.

I'm trying to figure out how this card can possibly be good; it doesn't seem very good to me, except its first ability. I suppose it combos with other dragons you have, as well as stuff like Day of the Dragons, but that's low-tier casual combo at best. Still, just because it's a planeswalker, I'd probably put it in all my 499-card decks. (I put one copy of all planeswalkers in those, except sometimes Nissa Revane, since she's so specific)

[Yggdrasil]

Thank you for these detailled statistics.

Highest Individual Deck Score: (just a cool statistic :p)
30 points - Yggdrasil, playing with Nyktorion's deck

Keep in mind that AI played 10 times with Nyktorion's deck during my gauntlets, it helped a lot (that and the fact it's clearly a good deck ).

If you play with custom version and want to include my vampire deck in your collection, please consider the following changes :
- 4 Guul Draz Vampire instead of the 4 Vampire Hexmage
- 2 Shriekmaw instead of the 2 Dark Depths

It was my first attempt before I realize that Guul Draz isn't available outside the custom version.
With these two changes, the deck is smoother and in my opinion a lot funnier to play. AI is also better.

[Yggdrasil]

As in the previous tournament, please feel free to submit me the card of your choice. I'll pick up one of them as my third card for the prize.

[aww1979]

When I am netdecking, the uncoded cards I run into the most are: (ignoring stuff I know will be uncodable, like Mutavault)

Trinket Mage
Squandered Resources
Powder Keg
Chalice of the Void
Pithing Needle (maybe uncodable)
Stroke of Genius
Future Sight
Verdant Force
Dwarven Miner
Necromancy
Skyshroud Elite
Replenish
Urborg, Tomb of Yawgmoth (might be uncodable)
Behemoth Sledge
Tithe
Dream Halls (probably uncodable)
Trygon Predator
Ethersworn Canonist (maybe uncodable)
Chainer's Edict
Counterbalance (though in manalink it would lose a lot of its power due to the old stack rules)
Kataki, War's Wage

I'm not saying these are the *best* cards, or even the most *fun* cards, but they seem to be the most *commonly used* uncoded cards that I find myself wishing we had. Except where noted, they all look easy to code and/or very similar to existing cards. Obviously, I can't request any of these, but I think that the short list above would go a long way to increasing the possible number of decks we can build, so I'd recommend any of these to the prizewinners.

@yggdrasil: If you take one of my suggestions, I'll say Future Sight. I *LOVE* that card.

[aww1979]

I forgot to deduct 5 points from those who didn't test. However, none of the rankings change due to this, except that shantak takes 8th spot and thedrigo drops to 9th.

[Yggdrasil]

I also didn't spend enought time building my own deck. If any of you have good ideas for changes or other ideas how to make a working Groundbreaker +Berserk deck, I'd be glad to hear them. What I've figured so far is:

-A bit more mana in some form. Though the spells are cheap, the chance of a bad opening hand seems too big.
-More destruction. 4 x Lightning Bolt might be a start, but I'm not sure what other spells would fit.
-Get rid of Spark Elementals. They were a relic from the time before Hellspark Elemental was in Manalink, and they really are a bit underwhelming in most cases.
-Maybe lose Forcemages? They're fun when their synergy works, but often little useless.
-Maybe splash White to enable tutoring for Oaths? Then again, like challenge mode has taught me, whenyou play with Oath, there are better creatures than Ball Lightning.

I like the idea of a deck revolving around Primal Forcemage, but it has several weaknesses :
- Primal Forcemage is fragile, especially in this competition where almost all decks are full of removals
- AI is unable to play it correctly

So I think it's probably a cool casual deck than a competitive deck for this kind of tournament.
However here are some ideas :
- Bloodbraid Elf seems to be a must have card (well, when the cascade's mecanism works correctly, of course )
- Viashino Sandstalker is also a very good creature
- Fists of the Anvil is strictly better than Blood Lust
- I am not totally convinced by the Oath of Druids in this deck
- Perhaps you should try with some direct damage spells instead of Giant Growth or/and Blood Lust ?
- Add some protection cards for your forcemages ? (Lightning Greaves, Tawnos's Coffin...)

[Salbei]

whow that was a surprise for me. i thought i would be in the mid field cause of the AI denying to play my deck^^

Already PMd Jatill my choice of cards i´d like to see.
Those are:

1:Extirpate
2:Urborg, Tomb of Yawgmoth
3:Chainers Edict
(yes, i like black)

Hope those are codeable. Consuming Vapors would also be nice to see.

[jatill]

aww-
Thanks for running the tournament. Those stats at the end were just fantastic!
Only one issue, though. According to the rules there will be a tie-breaker for 1st place by running the decks through another gauntlet to see which holds out the longest. As tournament director, I believe that would be your honor
Alas, that means I only get 4th place, thus no prize.

[Cognis]

Mutilate would also be neat, I also love black, zombies suck but I love them. How would have guessed we would have a draw. Congratulations to both Salbei & Yggdrasil although I'm a little surprised that a creature deck tied for 1st with so much removal in other decks, I considered vampires but didn't think it could work but the stats show that I'm a good player but suck at building

@aww1979: great job with the statistics, I love them, did you use something like SPSS or did you actually calculate it manually? Just one more thing, I lost to Tarmogeddon in the first round with my goblins so i went 0-1 (as I did last tournament with my own deck), not 10-0 I know the deck some mana issues but I though that 4 fetch lands, 4 Badlands and 5 swamps would be enough to go around Blood Moon and prove consistent mana but the problem was that all the black cards needed BB (except Demonic Tutor) which is a risk since one B mana is actually useless unless you have a second but this deck was made so the AI can win and the AI can't say no to landwalk. In the end I'm satisfied but I think I'll put in more effort next time in building.

[Cognis]

Alas, that means I only get 4th place, thus no prize.

Lol

[aww1979]

@jatill: Oops, I forgot about that... I don't really want to do that, though, hehe. If everyone else is fine with it, I think we should just leave it as two winners.

@cognis: I manually counted up the wins by each deck, and then Excel went nuts doing most of the rest And yes, you did go 0-1 with your own deck; I think I read the 10-0 off the next line in your list by mistake. (you went 10-0 with juzamjedi, and that is the line right under your own) Hopefully I didn't screw up anywhere else. I'll edit the post to correct that one.

[jatill]

@jatill: Oops, I forgot about that... I don't really want to do that, though, hehe. If everyone else is fine with it, I think we should just leave it as two winners.

Sorry, just 1 winner. Not gonna sucker me into coding extra cards Speaking of which, I won't have the chance to code the new cards until after the next tournament starts.