[NO LONGER A Premium Article] Investigating Skill

[Smmenen]

http://www.starcitygames.com/php/news/article/11983.html


The article talks about the components of skill and really highlights some of the huge pitfalls that people have to avoid when testing, when designing a deck, and when preparing for tournament.

I also explain what sort of decks you should be testing against, how to avoid straw manning an opponents deck in testing, explaining why non Drain, non Shop decks underperform, and all sort of other things. 

I think you guys will like it.  Speak.

[AJFirst]

Damn stupid SCG with their premium memberships and their...their Rock'm Sock'm Robots.

I'm getting a membership in the next month or so, and when I do I'll read it twice.
-AJ

[Moxlotus]

I realized one major point in the article not too long ago.  There is a reason we don't bother to test against decks like GrimLong and its exactly why you said.  I liked the article.  It was a good change of pace.

[sean1i0]

Yeah, it was definitely a nice article.  I also thought that ASL was a unique way of describing this subject with quite a bit of merit behind it.

[Smmenen]

What do you think is the most important implication of this article? 

[Moxlotus]

What do you think is the most important implication of this article? 

For me, it was your point on why you would be hard pressed to play a drain deck.  That and explaining why Drains perform the best and why its not necessarily because of the deck itself.

[Gekoratel]

Is there any reason that you didn't post an ASL for most of the major archetypes?  I realize that the values would be somewhat arbitrary but it's better than nothing.  I'm guessing that this is going to be covered in your next article.

[AngryPheldagrif]

In an article about skill and T1, I wish you would have gotten into the skill curve idea at least briefly; that being the idea that some decks are easy to learn but very hard to master, others being the opposite.

[Smmenen]

In an article about skill and T1, I wish you would have gotten into the skill curve idea at least briefly; that being the idea that some decks are easy to learn but very hard to master, others being the opposite.

Getrokal is asking a similar question.

I set out some factors for considering where and when decks are going to have higher or lower ASL.  Specifically, if they are more intuitive and less decision trees, then they are going to have higher ASL.  But the key is the relative importance of ASL.  That's why I came up with the concept.  Drain decks have higher ASL even though they are hard to pillot because of the collective expertise of the player base.  Decks like Grim Long and IT have lower ASL.  Decks that are mainstream have become the target of design hate.   Fish decks have a high ASL when played against mainstream decks.  The vertical distance between the top player and the average player is rather slim when piloted against Drains. 

[AngryPheldagrif]

For a more specific question: Many people dismiss Oath as far too 'easy' to play, yet time and again I see the same bad players scrubbing out with it and the same good players winning power. What is your personal opinion on the matter?

[Juggernaut GO]

For a more specific question: Many people dismiss Oath as far too 'easy' to play, yet time and again I see the same bad players scrubbing out with it and the same good players winning power. What is your personal opinion on the matter?

I would answer by saying: different metagames, and different pairings.

[Toad]

Long ago, when friends asked me for decks they should run at tournaments, I made a simple (and thus not extremely accurate) scale to make deck calls. It's based on 5 simple parameters. First one is the raw skills of the player out of 20. Second one is a modifier related to the skills of the player when It comes down to generic archetypes (Control, Aggro, Combo), as pluses or minuses ranging from -3 to +3. Third one is a modifier related to the skills level needed to pilot a deck, as pluses or minuses, from -5 to +5. Fourth one is related to metagame calls and tries to evaluate the strength of the deck according to the playfield, as pluses or minuses, from -5 (very hostile metagame) to +5 (metagame is a bye). Fifth parameter evaluates the experience of the player with the given deck, and the bonus is only used if they are higher than the one found on Parameter 2 (for example, a player who has never played Grim Long will get a 0 here, but some bonuses if he has some experience with Storm Combo already from parameter 2. If he has playtested Long intensively, he will get the bonus from #5, even if he is not really good with Combo in general), from 0 (never played) to +5 (master).

For example, you could have something like :

* Parameter 1
Player A, raw skills = 14.
Player B, raw skills = 8.

* Parameter 2
Player A with Control = 0 (neutral).
Player B with Control = +1 (OK).
Player A with Combo = -2 (very bad).
Player B with Combo = -1 (quite bad).

* Parameter 3 (examples, numbers are not accurate)
Oath = +4
Grim Long = -3
Slaver = +1
etc ...

* Parameter 4
Oath if Aggro metagame = +3
Oath if Control metagame = 0
Oath if Gifts metagame = -3
Grim Long if brainlessAggro metagame = +5
etc ...

This scale is quite accurate when It comes down to evaluate good decks to pick according to own skills and metagame. For example, It shows that It's better to play a deck we master a bit less in a good metagame than a deck we master perfectly in an hostile metagame.

Parameter 3 is the reason why Oath is performing properly. Bad players still scrub out since the +4 bonus (if we rate it +4) will not make them go higher than good players piloting harder decks to play. Nevertheless, the +4 bonus is enough to make average players as "skilled" as good players with decks with lower bonuses or even penalties. In the above example, without taking metagame (parameter 4) into account, player B would pilot Oath with a skills level of 13, while player A would pilot Grim Long with a skills of 9. Despite his low raw skills level, player B would still overall perform better than player A.

[Smmenen]

In an article about skill and T1, I wish you would have gotten into the skill curve idea at least briefly; that being the idea that some decks are easy to learn but very hard to master, others being the opposite.

This is implicit in the article.  Decks that have a low ASL tend to have a steeper skill curve.  Decks with a high ASL have a lower skill curve.

The importance of the ASL versus a skill curve concept is that a skill curve, in the objective sense, is irrellevant. What matters is a) where you fall on that curve (which SL modeling helps you figure out) and b) where your opponents fall, which the skill curve concept only suggests. 

But to be perfectly honest, I'm not sure there is a deck in Vintage worth playing that isn't very hard to master.

Diceman realized some time ago that knowledge and information is one of the most powerful things in vintage right now - as soon as a deck become identified as optimal, it loses some of its optimality.

Vintage is very recursive at the moment: solutions beget solutions.  If you play X, then I need Y or learn how to optimize my deck against X.  If you play Y, then you become vulnerable to Z, so I can run that.  etc.  If you want to fill in my symbols, just think Jester's Cap, Tormod's Crypt, etc and pretend you are using them in various matchups. 

I think I answered your other question in your Oath thread.

[Smmenen]

This article is no longer a premium article.


It's one of my favorites thatt I've written recently.  I thought this may be a good opportunity for non-premium members to check out what premium peopole get as well as to discuss some of the issues in the article with a broader audience.

Speak!

[Liam-K]

So you're saying if you're good at your deck, you will beat people who are bad at their deck?

Am I oversimplifying here?

[Dominik]

So you're saying if you're good at your deck, you will beat people who are bad at their deck?

Am I oversimplifying here?

No, he's saying that there are more factors that determine who has a better chance to win the matchup than just that. Another one of those factors is how well you know THEIR deck.

[Liam-K]

Yeah, knowing how to pilot your deck against the field is part of your skill with the deck.  As is knowing how to pilot your deck when you can't identify your opponent's strategy.  I guess it all seems a bit redundant to me.

[Smmenen]

So you're saying if you're good at your deck, you will beat people who are bad at their deck?

Am I oversimplifying here?

No, I'm saying that certain archetypes are inherently stronger in the hands of weaker players.   

[Liam-K]

So you're saying if you're good at your deck, you will beat people who are bad at their deck?

Am I oversimplifying here?

No, I'm saying that certain archetypes are inherently stronger in the hands of weaker players.   

Ok it seems worthwhile in that light.

[vroman]

This article makes plenty of sense. for example, I basically ignore the Dragon matchup, despite being relatively dificult for stax to handle. their key spells are at 2 and 3, so chalice is unreliable. welder or uba mask stops the decking version, but if they play a damage kill, thats irrelevent. however, Im just not very worried about Dragon, and dont concentrate sideboard or main deck construction with that deck in mind. bc the majority of Dragon players Ive faced tend to be subpar, and I dominate them. for some reason noobish players tend to enjoy Dragon combo. unfortunately then I get blown out by really good players who happen to pick up the deck, like Rich Shay at the gencon prelim.