[Article] Of Giant Growth, Burning Wish & Variance

[Jonathan Alexander]

Rather than talking about any specific deck, I tried myself on a more fundamental issue this week. While the content is not Legacy-specific, it is very important in deckbuilding, especially for linear decks.

If you want to see how these three things are connected, check out my article here:
Of Giant Growth, Burning Wish & Variance

Thoughts are appreciated as always!

[tsabo_tavoc]

Very well written! People want to win and have maximal control, while competitive players just focus on winning.

[twndomn]

Most often, this means we have decided on 55 or more cards in the maindeck, but our sideboard is still pretty much up in the air.

losing 5% in two important matchups will generally outweigh gaining 20% against a much less popular deck

Here’s another hypothetical scenario:We have a Deck X that wins 90% of the time if you play better than your opponent and always loses if you don’t.

What's up with all these numbers? Why 55? Why does it have to be 55 decided cards, not 56 or 54. What's the definition of an important match-up? Is an important match-up the same as most frequently occurred match-ups? Again, where do you get the numbers for your important match-ups will account for 5% and losing to an uncommon deck will account for 20%? These are arbitrary numbers with no mentioning of any statistical evidences.

The hypothetical scenario is pointless because it's never going to happen. If WotC screws up Standard, and one particular deck has the highest win rate, for example when Mirrodin came out and Affinity's win rate initially skyrocketed, it might be as high as 70% against every non-Affinity deck at one particular point in time. Then people just would give up and join the bandwagon, quickly Affinity's overall win rate will begin to drop because hate decks will become popular and there will be lots of mirrors matches for that deck. Hence, 90% win rate of a deck is not possible at any point in time.

I'm not asking you to be Steven D. Levitt, but here's a great article on how to use numbers in a magic article: http://www.hipstersofthecoast.com/20...game-analysis/

[TkDodo]

Nice read indeed, but isn't Krosan Grip actually good in RUG nowadays because of Rest in Peace ?

[Technics]

What's up with all these numbers? Why 55? Why does it have to be 55 decided cards, not 56 or 54. What's the definition of an important match-up? Is an important match-up the same as most frequently occurred match-ups? Again, where do you get the numbers for your important match-ups will account for 5% and losing to an uncommon deck will account for 20%? These are arbitrary numbers with no mentioning of any statistical evidences.

The hypothetical scenario is pointless because it's never going to happen. If WotC screws up Standard, and one particular deck has the highest win rate, for example when Mirrodin came out and Affinity's win rate initially skyrocketed, it might be as high as 70% against every non-Affinity deck at one particular point in time. Then people just would give up and join the bandwagon, quickly Affinity's overall win rate will begin to drop because hate decks will become popular and there will be lots of mirrors matches for that deck. Hence, 90% win rate of a deck is not possible at any point in time.

I'm not asking you to be Steven D. Levitt, but here's a great article on how to use numbers in a magic article: http://www.hipstersofthecoast.com/20...game-analysis/

Not sure if you're trolling or not...

First: If I am playing a deck, and I am better than the other player by a large margin (70% of the time I am) I SHOULD win 90% of the time. However I will only ACTUALLY win 63.5% of the time, because I am not better than 100% of players.

Second: The point is not the numbers, they are thrown out as examples, not facts. The point you SHOULD be taking away, is you need to LOOK at numbers, and make sure you are building decks based on numbers, and not "gut feelings". His article was a simple pretend the numbers are this, this and this. If that is true, then this. He is not claiming, or representing the numbers he grabs as being correct, but is telling you to go do your own homework, and USE numbers!

[Jonathan Alexander]

Not sure if you're trolling or not...

First: If I am playing a deck, and I am better than the other player by a large margin (70% of the time I am) I SHOULD win 90% of the time. However I will only ACTUALLY win 63.5% of the time, because I am not better than 100% of players.

Second: The point is not the numbers, they are thrown out as examples, not facts. The point you SHOULD be taking away, is you need to LOOK at numbers, and make sure you are building decks based on numbers, and not "gut feelings". His article was a simple pretend the numbers are this, this and this. If that is true, then this. He is not claiming, or representing the numbers he grabs as being correct, but is telling you to go do your own homework, and USE numbers!

Pretty much what I was going to say, yeah. Thank you.

Regarding the Deck X / Deck Y scenario, it is not actually that unlikely.
For example, look at Sam Black and Mono Blue Devotion right now. Sam's win percentage is way higher than everyone elses. Clearly, this has to do with his playskill. Or look at the Mono Black list from GP ABQ. Their performance was insane.
The same was true multiple times in the past when Team CFB had an average constructed match win percentage of ~70% at a PT, while the deck they were playing was much closer to 50% overall.