What I Learned Yesterday #14

It’s not a good idea to postpone things. Seriously. Now it’s much harder to remember what information I’ve got yesterday is worth sharing… But here we go:

  1. Facebook shares dropped in price and its a very good time to buy it. This company will stay on market for a long-long time and will always find ways to get profits. Mark knows how to do that.
  2. Now I have an idea how to put machine learning model in production. I mean working idea. You create your model in python, then you create an API using Flask library, then you put it into production using Docker – and it’s live. Easy to say, not that easy to implement. But I have a course on this so I will find a way and probably will create something to play around.
  3. Something about logical thinking. Just a little story about broker’s scam (didn’t happen in reality). Imagine one day you receive an email from unknown address who claims that shares of some companies will go up next week. You definitely ignore this guy, but the next week you see that those shares really go up. Next Saturday you receive an email again saying that other shares will go down and they really do. And the same repeats every week and every time the guy is correct in his predictions. On the 10th email he offers you his help in investing your money with really high commission (he proved his credibility, didn’t he). And you think: “Oh, gosh, he was correct 10 times in a row, this guy should know what he’s doing”. And it really makes sense. But let me tell you the same story from the broker’s perspective. On week one the broker sends 10540 emails: half of them with the prediction that shares will go up and the other half with opposite forecast. He waits the next week and sends the same, but only to those 5270 emails which received “correct prediction”. At the week 10 there will remain 10 email addresses which received 10 “correct predictions” in a row and they will definitely hire the guy. Do you see the irony? Past results don’t indicate the future’s, especially when we deal with probability. So when you see something improbable but possible happening, ask yourself: “How many tries this event had previously?”. I am not telling that improbable thing don’t happen, they actually do and happen a lot, but we have to be critical, especially if we talk about money 😉

More stuff similar to the broker’s story you can find in amazing book by Jordan Ellenberg “How Not to Be Wrong: The Hidden Maths of Everyday Life”. Here is a link to Amazon – https://amzn.to/2LXoLd5

Hopefully it was useful for you, my reader, have a nice Saturday and be happy. Today’s quote will be about probability, it is really a fun and exciting thing.

Leave a Reply

Your email address will not be published. Required fields are marked *