Author Topic: Monte Carlo  (Read 2641 times)

Warrior_Monk

  • Guest
  • Trade Count: (0)
Monte Carlo
« on: March 23, 2015, 01:35:48 PM »
0
Rather than going back and actually having to remember how to calculate probabilities, I wrote a quick program that simulated drawing over 1,000,000 trials or so.
Me too!  (Even the same number of trials. ;) )

Hey, Kids, here is a pro-tip for you...Markov chains are much easier in Monte Carlo.
So this was recently brought up in another thread, and the idea has been something on my mind for awhile now, but I don't know where to start.

Suppose I wanted to create a monte carlo where I have a table of deck archetypes, each with matchups against each other, and each with percentage of people playing it, and the number of people (so it can automatically calculate number of people playing each deck but I can modify percentages as I see fit). Then I have a number of rounds.  What is the best way of setting something like this up?  Does anything like this already exist?

Simple Example:
MatchupsRockPaperScissors
Rock50%0%100%
Paper100%50%0%
Scissors0%100%50%

Frequency
Rock40%
Paper35%
Scissors25%

Number of People:345
Number of Rounds:9

Offline EmJayBee83

  • Tournament Host
  • Trade Count: (+1)
  • *****
  • Posts: 5483
  • Ha! It's funny because the squirrel gets dead.
    • -
    • East Central Region
    • mjb Games
Re: Monte Carlo
« Reply #1 on: March 23, 2015, 02:06:49 PM »
0
I don't know of anything simple that would do that for you. (That doesn't mean there isn't one--I just don't know of one.) I  If you know Excel, MATLAB, or any programming language this would be really east to set up.

Warrior_Monk

  • Guest
  • Trade Count: (0)
Re: Monte Carlo
« Reply #2 on: March 23, 2015, 02:24:56 PM »
0
I only sorta know Excel. I guess I'll do some googling and see if I can figure it out.  I had previously read before on it, but it didn't do everything I wanted it to do.

Additionally, I'd ideally like to see not only the winner, but also the average number of said deck in Top 32.

Rawrlolsauce!

  • Guest
  • Trade Count: (0)
Re: Monte Carlo
« Reply #3 on: March 23, 2015, 02:37:29 PM »
0
This doesn't seem too difficult to do with Matlab. http://www.mathworks.com/discovery/monte-carlo-simulation.html

IDK if you have Matlab. I know I get 1 copy for free from my university, so maybe it's the same for you.

Offline Alex_Olijar

  • 16plus
  • Trade Count: (+15)
  • *
  • Posts: 8124
  • This guy is my mascot
    • -
    • Northeast Region
Re: Monte Carlo
« Reply #4 on: March 23, 2015, 04:03:33 PM »
+1
this is just simple monte carlo stuff, had to do this for tons of business classes. go to real school not rock school and you could do it too.

Warrior_Monk

  • Guest
  • Trade Count: (0)
Re: Monte Carlo
« Reply #5 on: March 23, 2015, 04:52:54 PM »
0
this is just simple monte carlo stuff, had to do this for tons of business classes. go to real school not rock school and you could do it too.
Should I edit OP to say "plz do it for me I'm lazy"?

Offline Alex_Olijar

  • 16plus
  • Trade Count: (+15)
  • *
  • Posts: 8124
  • This guy is my mascot
    • -
    • Northeast Region
Re: Monte Carlo
« Reply #6 on: March 23, 2015, 04:55:23 PM »
0
yeah kinda

Offline kram1138

  • Trade Count: (0)
  • Full Member
  • ***
  • Posts: 431
    • -
    • North Central Region
Re: Monte Carlo
« Reply #7 on: March 23, 2015, 06:50:31 PM »
0
If you're more precise about what you want it to do, and If I have time (which is a big if), I might be able to whip something up that could do it.
postCount.Add(1);

Offline Alex_Olijar

  • 16plus
  • Trade Count: (+15)
  • *
  • Posts: 8124
  • This guy is my mascot
    • -
    • Northeast Region
Re: Monte Carlo
« Reply #8 on: March 23, 2015, 09:11:54 PM »
0
he essentially wants to run a 250 person tournament in excel using winning percentages of deck matchups against each other, random pairings, and determine how likely a specific deck variant is to win the tournament (in swiss)

Offline kram1138

  • Trade Count: (0)
  • Full Member
  • ***
  • Posts: 431
    • -
    • North Central Region
Re: Monte Carlo
« Reply #9 on: March 23, 2015, 11:33:50 PM »
0
That's sounds pretty straight forward, but I don't know how to use excel. I could program that pretty easily though.
postCount.Add(1);

Warrior_Monk

  • Guest
  • Trade Count: (0)
Re: Monte Carlo
« Reply #10 on: March 23, 2015, 11:46:13 PM »
0
That's sounds pretty straight forward, but I don't know how to use excel. I could program that pretty easily though.
I mean, as long as I can use it, I don't care. I was just going to try excel because that's what I have available to me.

Offline Alex_Olijar

  • 16plus
  • Trade Count: (+15)
  • *
  • Posts: 8124
  • This guy is my mascot
    • -
    • Northeast Region
Re: Monte Carlo
« Reply #11 on: March 23, 2015, 11:49:39 PM »
0
variables we can enter should be number of people, number of decks, individual decks win percentage against other decks, frequency of each deck in the field, and number of rounds. i think that's it.

Warrior_Monk

  • Guest
  • Trade Count: (0)
Re: Monte Carlo
« Reply #12 on: March 23, 2015, 11:52:27 PM »
0
I may have convinced Olijar to get on board with this idea.

 


SimplePortal 2.3.3 © 2008-2010, SimplePortal