I am a poor noob without much experience in trading, but with some know-how in computer programming. After many years of hard working I was able to save $3k. In fact, a little more than that, but that's what I am willing to risk. I know you elite traders are all millionairs and will tell me that there is no place under sun for us poor pikers. The purpose of this thread is to prove the opposite. I made some simulations and found an intra day trading system for US stocks that looks profitable. The system can be tuned to select only most profitable signals to return on average 1.5%/trade, but frequency will be only one signal per day. At the level of 0.8%/trade there will be ~5 signals a day. I'll start with larger average return setup, and will gradually move to larger frequency setup as my capital grows. As everyone knows, there is a PDT rule that says: if you have less than $25k in your account, you must hold your shares overnight. You cant sell shares on the same day as you bought them. Luckily, this applies only to US brokers. So I am going to open an account with questrade.com, a canadian broker and I can trade intraday as much as I want.
Applied to the broker online, received some agreements to sign, and they requested a notarized copy of my passport. Printed, signed and sent them by mail. Now waiting for confirmation of account opening before wiring my $3k to them.
While waiting for paperwork to clear, continue to double-check the system. Split adjustments are pain in the ass. I store end of day quotes in a database, and there is an "adjclose" column to account for splits / dividents. Real time quotes are in plain binary files, not adjusted. If I want to run simulations for some past date, I should reconstruct EOD quotes for that date, so that they match quotes in RT files.
Some backtesting results on 99-days period Sep06-Feb07: conf 3.13 conf1 3.36 av% 0.94 n 103 +68 -35 result 97% Where "conf" and "conf1" are confidence values, computed by two slightly different formula. Approximately, it is -log10(propability to achieve that result randomly). In fact, some trades are correlated, so these confidences are only rough indication. av% is average percentage gain on trade, n is number of signals, + is number of winners, - is number of losers, result = n * av%. The big unknown are actual fills that I will be getting. Currently I use this method in simulations: entry/exit price is average ask/bid price within the next 30 seconds. I consider stocks with average daily volume > 200k. That's considerably worse than preliminary results based on only 31-days period sep06-Oct06. Probably because stocks rose during that time. Should try to improve the strategy.
Tweaked a system a liitle bit, the result over 99 days: conf 13.13 conf1 6.07 av% 1.60 n 215 +161 -54 result 343% Decent average gain, frequency about 2 signals per day, so majority of days have at least one signal, that's what we look for. Can we improve more?
Hey smart guy. Take a long hard look at the e-mini futures (ES for S&P500, NQ for the NASDAQ, YM for the Dow and ER for the Russell). Study them and study them hard. If you can crack their code the way you have with stocks, you'll clean-up. Good trading, Jimmy Jam
Thanks, but haven't cracked anything yet - that's all paper dreams ... Got about 300k 1-minute bars for ES. Would love to have more data, and at higher resolution. Anyone willing to share ?
any updates? Im in the UK and am thinking of opening an account with questrade as a way of getting round the pdt rule?
No luck so far. Initially they approved my account and told me to send funds. I did so, and then they revoked the approval. Now my money sits in the account but I cannot trade. They keep asking for a check from Canadian bank in Canadian currency. I explain that have no account with any Canadian bank, then they ask my passport to be certified by Canadian lawyer. I explain that I have no access to Canadian lawyers, then they don't know what to do. I call again, another representative on line, all starts over again. Apparently not much foreigners register to trade through them, though their Web site seems to accept applications from overseas. I'm not in a hurry however, because I decided to download historical tick database from OpenTick, to test my system on more data and on larger time frame. OpenTick is essentially free (I registered two accounts for $4 a month), so speed of data retrieval is .. you get what you paid for I estimate it will take me another two weeks to complete downloading historical data.
I guess if I do not receive approval within a month, I'd request my money back and try alliancetrader.com. All that situation where they took the money and then started to play approval games, is very wierd. Hope I did not fell victim to Web criminals.