Welcome to Draw 3 Cards

Question

2

I don't understand why some people think this card is so good.

Lands are about 45% of most decks. So (without any library manipulation techniques) you have:

55% chance of drawing one card for 2 mana
25% chance of drawing one nonland card & one land
11% chance of drawing two nonlands and one land
5% chance of drawing three nonlands and one land
...

The expected (= average) number of nonland cards you'll get is about 0.8 cards, if I'm not mistaken:

1-1/(0.55) = 0.8

So, looking at it like this, 2 mana for an average of 1.8 cards is not bad. However, the average return is not the whole story. After all, if a spell would give you 0 cards 90% of the times and 20 cards 10% of the times, it wouldn't be very good, even though the average number of cards it draws is 2. Most (55%) of the time, this card simply wastes two mana to get another card which will not be a land. In these cases, it won't fix your mana base, which is something I usually rely on card-drawing spells for.

So my question, after this analysis, is what kind of deck would you use Treasure Hunt in Limited, and when would it rest in the sideboard?

[[Treasure Hunt]]

asked Feb 5 at 15:32

ripper234♦♦

community wiki

Shushoto · Answer 1 · 2010-02-12 04:49:43Z UTC

Actually, the specific average draw from Treasure Hunt is modeled by the series:

(X-Z)n*Z!(X-n)!/(X!(Z+1-n)!), 1<=n<=Z+1

Where X is the number of cards left in the library as Treasure Hunt resolves and Z is the number of lands left in the library.

I have a partial series program on my trusty TI-83+ so I'll give the numbers a whirl and:

45% land with 30 cards left gives 1.771 on average, with 20 cards gives 1.750, etc. (if there were infinite cards in the deck with that ratio then yes 1.8 would be average but the smaller the deck gets, the more inaccurate that number)

On the other hand, adjusting the percentage of land in the deck (these are exact data points, but will follow the trend shown above if deck size is modified):

     80% land, 30 cards - 4.429
     70% land, 30 cards - 3.100
     60% land, 30 cards - 2.385
     50% land, 30 cards - 1.938

As you can see, a larger percentage of land has a great impact on the average draw ratio (lowering the land ratio should just lower the average to 1).

I know this doesn't directly address your question, but I'm a mathematician at heart so I couldn't help myself. Hmm, probably should have put the apology in advance, are you still reading this?

My speculation is that Treasure Hunt's strong point (unless used for combo) is to ensure a non-land draw. If it gets extra land you need for landfall, then great but it's ultimately a deck thinning card in my opinion.

Krinn · Answer 2 · 2010-02-07 01:30:52Z UTC

Well, the very simple answer is that you'd use it in a deck that was landfall-heavy - where the EV was more than 1.8 cards. If you had a high enough concentration of landfall cards and lands with enters-play effects, I could see a deck that had 18-21 lands. That deck would love Treasure Hunt, because it would be worth more cards and would ensure that you'd always have a way to trigger crucial landfall effects.

Whether that deck would be good or not, I will not comment on.

mason · Answer 3 · 2010-03-01 05:47:27Z UTC

2

[[Halimar Depths]][[Jace, the Mind Sculptor]][[Sphinx of Jwar Isle]]

Treasure Hunt is ok because it always draws a spell. It is extremely good in combination with any of the three cards mentioned here, all of which are great cards on their own.

answered Mar 1 at 5:47

mason

community wiki

DarkMantle · Answer 4 · 2010-03-02 03:50:04Z UTC

1

If you put in Selective Memory and pulled cards that you don't want/need you would increase the odds of pulling lots of land.

[[Selective Memory]]

answered Mar 2 at 3:50

DarkMantle

community wiki

Where would you use Treasure Hunt in Limited?

Remember to vote up questions/answers you find interesting or helpful (requires 15 reputation points)

4 Answers