First off, I design all my own ideas when it comes to formulas. I may use an existing 'out of the can' indicator, but I never use them as they were designed, I invariably always twist an 'out of the can' indicator into something else. I have little faith in indicators as such but my opinion is they can have uses in other ways. I'll try and explain... First off I'm looking to buy quality stocks. I identify them largely, initially by using T/A, by low volatility and the fact they run consistently, trend consistently, don't plunge often suddenly etc. Many penny dreadful type stocks are best shorted I think, they are not good for consistency heading upward. You can nearly always guarantee a crap stock will drop off a cliff. The inverse being true, good stocks rarely drop off a cliff, but there are exceptions. Also, quality stocks wont be in extended downtrends, but there are exceptions. I trade probabilities, so those which are the exception, too bad, they get dumped from my watchlist regardless. To find good trending stocks and low volatility I look primarily long term over the full life of the stock. A new issue may only have been trading several months, so that becomes how long the lookback is. Note: New issues are often the best contenders, after about 4-5 years they don't appear to perform so well, as a general rule. Anyhow, my selections are done from a formula exploration, ie I run the exploration which pops up a list of contenders. The exploration formula has a section with the lookback period and counts how many bars the stock has traded. Another part looks at bars and identifies them as medium height, ie I don't always read close or open, but medium height. Another section has filters for including or excluding contenders, eg, if it trades infrequently, is a warrant or option, if over the life of the stock it's trended more downward than upward, and identifies those highly volatile, only traded a few days, is below 10cents etc etc The meat of the formula looks at price action, ie what has the stock done in terms of how price/bars behaves. I'm not too much into volume, i've found volume short term gives too many conflicting signals, is unreliable, a red herring, distracting, confusing. Price action is read looking at volatility, pullbacks, trending, how often hitting short term supports and resistance, ratios of gapping up and down, how often price collapses on volume spikes etc I run this bunch of key things via the formulas, then i score them. I score them long, medium and short term. The long medium and short term scores are kept separate, they are there for me to pick which time frame i want, but mostly I look at long term. Some criteria are only for short term, ie, the gapping ratio is a short term thing, it is not part of the long term score. Anyhow, that's the basic nuts and bolts of it. In answer to nutmeg's question, I've come to 40 positions as it spreads my risk to approx 2.5% per position. If one position completely goes out of business and i loose everything, that's 2.5% of my portfolio. I've found running 20 positions at the moment no effort at all. As I probably only buy or sell maybe once every 2-3 weeks, this is very easy to manage. Amibroker is very friendly when it comes to tracking positions. I run a large spreadsheet for taxation purposes, I've used the same spreadsheet unaltered in terms of structure for several years and my accountant and i have no trouble quickly getting off it the end of years results for all my trading which includes the dividends which may fall into my lap.
A pointer to how i may twist an indicator. For example I may use RSI, but instead of reading RSI directly, I will measure how often a RSI hits 20 and how often it hits 80, count the ratio over a lookbackperiod and this may tell me whether ABC stock primarily surges upward or surges downward. If you compare this ratio over numerous stocks, one could eliminate the baddies and keep the goodies. Do this using several different criteria and that's how you can build up a score of best and worse contenders That's about it, probably be reluctant to give away too many further ideas, if probing questions on my system are requested, i may not divulge. My primary task was to show that a long term trading system has merit, I'm not wanting to divulge too much of the ingredients. So far the questions have been good. Thanks
I had worked out a similiar idea a while back. My intention was to buy (to use your example $2500 of one stock). Now if the stock went up 10 or 15 % I would sell shares equal to the profit so my remainder balance was back to $2500. Then use the profit plus whatever additional money that would equal the $2500 starting point and purchase another stock. The losers I would treat as such and follow whatever course of action I used to exit losing trades. I set a predetermined amount I would use to purchase as many stocks till I was at my max drawdown, total of wins and losses.
I don't have in mind any particular annual percentage gain, just take what is offered. The attempt is to compound. ie, I can loose $2500 if the stock goes to zero but gain unlimited should they go up and continue climbing. Premature scalping a profit is not what I wish to do, the theory being, if I take care in my selection, probabilities are the stock will rise in line or exceed market rate. Selling is only done if I see historical turnover getting excessive, then I use a monthly stop. When I enter, should the stock go down, I don't exit, I just wait for it to (a) go to zero or (b) recover. Hopefully (a) won't occur, but if it does, that's the price I pay.
What is your typical hold time? My typical hold time is only a few days but I also trade a couple long strategies that typically hold a few weeks and I have been working on some strategies that will be easier to trade and I'm starting to get hold times measured in months now. Seems like forever compared to what I've been trading but also much easier to do and more scalable. I agree with the comment that the drawback to this approach is likely to be the drawdowns. If you can identify stocks likely to go up you should be able to identify stocks likely to go down and then you can trade long and short.
Still early days for me to be able to able to comment on typical hold times, but looking at my present positions I'm guessing a minimum 6 months, max 3-4 years, per position. Drawdowns shouldn't be too vicious unless the whole market tanks, then it's just a matter of marking time until market recovery. With 40 positions, there hopefully will be a smoothing effect. I look also for those which are key industries, civil engineering, medical, University education, energy - oil, gas, uranium, car sales, internet service providers, materials, new technology etc. The tactic is choose very carefully the selection so as to increase odds of price increases. Just a calculated gamble, no different from any other type of trading, but having the luxury of slowing down the whole process so one has time to weigh up the odds. I'm careful not to buy into stocks which look like they have already run hard, i attempt of find stocks which meet my criteria and have relatively new monthly buy signals.
Another couple of reasons for going this way... There seems to be a race on toward faster and faster Hi.Freq.Trading, I can't compete in this environment so while a mass of traders go short term, i decided to head in the opposite direction which isn't so crowded. As well, I'm weary of doing homework on stocks, buying in, getting stopped out, then months later reviewing and seeing these positions still running up but i've missed the opportunity due to distraction, I'm all over the place due to lack of attention on having an eye on the ball.
I am long term trader too. I try and ride the momentum whichever way it goes. From March to December, I only had 21 trades.
I also like this strategy but I would hesitate implementing it. I recall it was very popular in the late 1990s. many with investments in dot.com bubbles never recouped their money. The key question is about the stop-loss. If all stocks start going down in synch when do you say this is it and you get you?