I re-read Michael Lewis' Liar's Poker once every few years. In new reprints he's added a prologue where he says something along the lines of "I tried to write

I re-read Michael Lewis' Liar's Poker once every few years. In new reprints he's added a prologue where he says something along the lines of "I tried to write a book about the horrors of wall street and instead it became a recruiting tool for wall street" In a similar vein whenever something like this comes up I write something that says don't try and do this and inevitable I get 10+ emails from people saying "Ok I understand but I still want to do it anyway." So if you really want to try and day trade, here is my 5 points of advice 1) You haven't done anything until you trade with real money. Ever played poker with friends that wasn't for money? You always get some jerk bluffing and going all in on Ace-7 because why not, it doesn't matter. If you want to trade you need to learn your own emotional limits. Get invested as soon as possible. Learn what kind of stress you can deal with. 2) The average successful individual algo trade looks like John Turturro charter Joey Knish in rounders. You are always grinding out a living $100 to $1000 at a time. If that isn't' appealing then stop before you start. http://www.imdb.com/title/tt0128442/fullcredits/ 3) You need money to begin. It takes as much effort to trade 2000 as it does 200,000. Above a million or so things change. A minimum amount I'd use would be 50,000. If you can't set that aside to trade then don't start. I mean if you have a great year trading 5,000 and make 20% you have made $1,000. That's not worth the effort. But at say 100,000 your 20% is starting to be real money!! Be clear to yourself that you are doing this for the money. Any other reason just gives you an easy out. 4) The most important number to look at is draw down, not sharp ratio. I see lost of people on quantopian show great backtested algos that return 200% over 5 years, but have a 70% draw down at one point. Most people don't have the psychology to loose money. At 10% down you start to question yourself. At 20% down you get really grumpy in all aspects of your life. Most people don't have the ability to let their positions fall to 30% down. If your backtest has a draw down of more than 30% you don't have a workable algo.....period. Sharpe ratio is great if you are an HFT firm trying to decide how to divvy up your money across 30 successful algos. You don't have 30 successful aglo's. You will be lucky to have one. 5) Don't use leverage....... you can break the other rules, this one you can't. I've seen people go bankrupt. Don't use leverage.