Archive for March, 2019

Warning: Use of undefined constant archives - assumed 'archives' (this will throw an Error in a future version of PHP) in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32

Warning: Use of undefined constant page - assumed 'page' (this will throw an Error in a future version of PHP) in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32

Warning: A non-numeric value encountered in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32

Warning: A non-numeric value encountered in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32
class="post-7733 post type-post status-publish format-standard hentry category-technology tag-math tag-monty-hall tag-programming tag-python">

Monty Python Hall: A Monty Hall simulator in Python

March 25th, 2019


Public domain illustration from Wikipedia
 

The infamous Monty Hall problem goes something like this: Monty Hall, the host of TV game show “Let’s Make A Deal” shows a contestant three doors. Behind one door is a new car. If the contestant picks that door, they win the car.

The other two doors have a goat behind them, which are not prizes. I guess this puzzle falls apart if you for some reason find a goat more valuable than a car. Anyway, point is the door with the car behind it is the prize.

Once the contestant picks a door, the host Mr. Hall opens a second door to reveal a goat. Now Mr. Hall asks the contestant to select one of two options:

  1. Stay with the original door they picked
  2. Switch to a different door (the one remaining door that is not open)

Here’s where most of us get tripped up: it shouldn’t matter, right? There are only two remaining doors, one has a car behind it and the other a goat. Why does it matter if you stay or switch doors?
 

What’s going on?

As it turns out the odds in this situation are counterintuitively not 50%. When the contestant initially selected a door the odds they selected the prize door were one out of three. But this changes once Mr. Hall opens a door to reveal a goat using his prior knowledge. We know Mr. Hall will never open a door with a car behind it, and he will never open the same door the contestant initially selected.

The contestant’s initial choice remains fixed in time: one out of three. Now that a door revealing a goat has been opened the initial choice is still one out of three. But if the contestant switches doors, something unexpected happens — their new choice has a two out of three chance of winning.

If this problem sounds familiar you probably heard about it in Parade magazine back in the 90’s when columnist Marilyn vos Savant spent years covering this problem. It’s also been tested on Mythbusters. For a quick explanation check out AsapSCIENCE’s video on YouTube.
 

Another approach

Like many people, I often find probability mathematics baffling — and the Monty Hall problem is no exception.

When I was a kid I remember visiting The Exploratorium here in San Francisco and coming across a simple computer exhibit that simulated dice rolls. I’d never considered before that when you roll one six sided die the probability of landing on each side is equal, but when you roll two dice and add the numbers together, the probability of rolling any number between two and twelve are not equal. There’s only one way to roll a two or a twelve, but plenty of ways of getting, for example, an eight.

The exhibit drew a graph of the number of times it simulated a two-dice roll, with a bar for each possible outcome. After some number of rolls it always looked like a hill with most of the rolls in the middle range and fewer at the higher and lower numbers.

Since I was learning to program at the time, this idea of taking a mathematical problem and breaking it down into a computer simulation really appealed to me. In fact when I got home I wrote my own version of the dice roll simulator in QBasic. In only an hour or two I managed to put together a working dice roll simulator complete with a bar graph, just like the exhibit at The Exploratorium.

A computer simulation isn’t a mathematical proof of course, but it’s a good sanity check to validate a hypothesis.

Thinking about the Monty Hall problem again recently I thought I’d take that approach I’d learned as a kid and write a simple program to simulate it and tally up the results.
 

Monty Hall problem in Python

As a software engineer I mostly work in Python these days. It’s a fairly easy to understand language so I thought it’d be perfect for simulating the Monty Hall problem. That’s how I came up with “Monty Python Hall,” a Monty Hall problem simulator in Python. The idea is to run the Monty Hall three door problem any number of times and tally up the results at the end.

Once you’ve installed a Python interpreter on your system you can try out my Monty Hall simulator from the command line. Clone the repo, open the directory in a console and type “python run.py” and you’ll see an output like this:

Games run: 1000
Games won stayed: 351
Games won switched: 683

If you edit the source you can change the number of games run, but the result always comes out about the same: the contestant wins 2/3 of the time if they switch, and only 1/3 of the time if they stay.

I designed the program to prioritize simplicity over efficiency, so if you run it too many times it may be slower than you expect. For example look at the way the host selects a door:

while host_choice == player_choice or host_choice == car_at:
    host_choice = pick_random_door(num_doors)

This is more complex than it needs to be; the host’s choice is simulated randomly until it meets the required conditions.

Why? I think it’s more interesting to let people play around with the program, and the simpler the logic the easier it is to modify.

For example what if you change the number of doors? It needs to be at least three for the contestant and host’s choice to work. But what if there were ten doors? In my version of the program the host opens one door with a goat behind it no matter how many doors there are. But what if the host opens half the doors with goats behind them? Or all of the doors except the one that may hide a car?

My implementation of the Monty Hall simulator in Python is available under the free, open source MIT license. I encourage you to try it and modify it as you see fit.

If you find this useful in any way feel free to send me an email. I’d love to hear about it!
 

Find this project on Github

Warning: Use of undefined constant archives - assumed 'archives' (this will throw an Error in a future version of PHP) in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32

Warning: Use of undefined constant page - assumed 'page' (this will throw an Error in a future version of PHP) in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32

Warning: A non-numeric value encountered in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32

Warning: A non-numeric value encountered in /home/public/blog/wp-content/themes/eric-cordobo-green-park-2/archive.php on line 32
class="post-7708 post type-post status-publish format-standard hentry category-local tag-art tag-civic-center tag-photos">

W.F.T (San Francisco)

March 14th, 2019

W.F.T (San Francisco)
W.F.T (San Francisco) W.F.T (San Francisco)
 

This evening I decided to take a walk by the Civic Center area to check out a brand new art installation: W.F.T (San Francisco) from artist Joseph Kosuth. W.F.T. — or “Word Family Tree” — is a neon light piece that wraps around the Polk St. side of the Bill Graham Civic, lighting up an otherwise boring brick wall with bricked-up windows.

I was held up too late at work to make it to the lighting ceremony though by the time I arrived there were still a few people lingering around taking photos. From the street level it’s easy to miss; the neon lights seemed dimmer than I’d expected, and are high enough from the ground level that the best view is from across the street.

The neon lights form trees breaking down the etymologies of the words “Civic” and “Auditorium.” It almost looks like notes taken by a college student in a linguistics course, except if those notes were inexplicably turned into light and attached to the wall of a four story building.

While I admire the unusual decision to put a brainstorm cloud of words in neon on the side of the building, the unfortunately ugly fire escapes get in the way, literally blocking your view depending on where you stand. It’s hardly the fault of the piece though.

My only critique is the use of the word “auditorium.” To most of us locals the building is called “Bill Graham Civic.” Honestly I’d forgotten the full is name of the place is “Bill Graham Civic Auditorium” until today. It’s not ear splitting like when an out-of-towner says “The BART,” but if someone told me to meet them at “the auditorium,” I wouldn’t know what they were talking about.

The building itself dates back to 1915 when it opened as part of the Panama Pacific Exhibition. Since then it’s served various purposes, including a basketball arena for the Warriors, an opera house for San Francisco Opera, and an exhibit hall where a very early prototype of what we now think of as a computer was first demonstrated.

Also, I saw The Smashing Pumpkins play there once back in ’98. Cool show, man.

These days it’s primary a concert venue. The name was changed by the city in the early 90’s to honor legendary local concert promoter Bill Graham after he died in a helicopter crash.

If you’re in the area when it’s dark enough — whether for a concert or just getting off work — you’d might as well wander by the Polk St. side of the building and take a look at W.F.T. for yourself. Until now you probably haven’t seen neon signs written in Greek and Latin, let alone many of them stuck all over the side of a building. There’s plenty of time to check it out in person as it’s considered a permanent installation. Of course even in the best conditions neon lights don’t last forever; better to check it out sooner than later if you’re interested.