Traditional pyramiding on your initial position tells you to adds to your position in smaller and smaller increments. But look what happens when I add to a position in a Fibonacci sequence on the ES every 20 points along with a Stop placed at 20 points below the last purchase: Qty Bot, Price Bot, Avg Cost, Stop, Profit 1, 1400, 1400, 1380, ($1,000) 1, 1420, 1410, 1400, ($1,000) 2, 1440, 1425, 1420, ($1,000) 3, 1460, 1440, 1440, 0 5, 1480, 1456.67, 1460, $2,000 8, 1500, 1474, 1480, $6,000 13, 1520, 1492.12, 1500, $13,000 21, 1540, 1510.74, 1520, $25,000 34, 1560, 1529.77, 1540, $45,000 55, 1580, 1549.09, 1560, $78,000 89, 1600, 1568.62, 1580, $132,000 144, 1620, 1588.30, 1600, $220,000 233, 1640, 1608.08, 1620, $363,000 377, 1660, 1627.93, 1640, $595,000 610, 1680, 1647.83, 1660, $971,000 987, 1700, 1667.77, 1680, $1,580,000 1597, 1720, 1687.72, 1700, $2,566,000 You could even modify the sequence a bit so that instead of (1,1,2,3,5) you add a 1,0 to the front so it becomes (1,0,1,1,2,3,5). By doing so, you reduce your losses from the first 3 levels to only the first level. Obviously there are some practical problems. What volume can the ES handle during the less liquid night session. Plus you could have big gaps between the close at 15:15 Central time and the reopen at 15:30 Central time. An even bigger problem would be the a gap against you between the close on Friday afternoon and the open on Sunday. I think I could live with the 15 minute gap during the week but the long time between Friday and Sunday would make me want to close my position on Friday and reestablish it on Sunday. Certainly you'd want to look at the 20 point step level and chose the optimal step value. You could also use this as a simple strategy that is always either long or short the market. When you get stopped out on your long position, go short 1 ES at the same level you got out of your long and if the ES keeps going down, pyramid up in Fibonacci fashion on the way down. Anybody with backtesting software care to run this through your software to see what you get? What's the biggest position you ever get to? How often are you stopped out on your first purchase, 50% of the time? Is this just pie in the sky stuff or are there some fundamental rules (let your winners run, cut your losses short) being exhibited here?
I worked on something similar that was based on standard deviations from a moving average, buying more as price moved away. It tested out as a good strategy with the caveat that every so often you would blow out your account and have to start over. It became an odd thing to think about really, overall profitability was very much affected by how often and how badly it blew up the account.. I never completed the work but it was a fun thing to play with for awhile... Possibly a lot of these blown up hedge funds had similar ideas and simply traded it with other people's money and collected big fees until it blew out, seems likely...
Did your system buy more as the price moved up or buy more when price moved down? Since my strategy only losses $1,000 per trade (and less if you use a step size less than 20 points) it seems like it would take a while to blow out a decent sized account with my system.
You forgetting you need the ES to move 300 points. Also when you are loaded up it will take just one bad day to wipe you out and that will happen frequently. You should cap the martingale after 5 or 6 increases.
The idea is to place the stop right after you buy at each level. For example on this level 13, 1520, 1492.12, 1500, $13,000 I buy 13 ES at 1520 which makes my average cost 1492.1212 after buying the 13 contracts. I then place an order to Sell the full position 33 at 1500. If I'm stopped out, I've got a profit of 7.8788 * 33 * 50 = $13,000.02. So a bad day will just hit my stop but shouldn't cause the trade to turn into a losing trade. Your point of the ES needing to move 300 points without a 20 point retracement is a good one. But I'd be interested to see what percentages each level is reached in backtesting this strategy. You'd still have to handle potential liquidity issues and gaps for the times between the ES closing and reopening.
Instead of using a 20 point step level you could use a smaller level, say 5 point increments. That would reduce the amount that the ES needs to move in your direction. This is an Anti-Martingale system in that the position is added to only if the ES moves in my direction.
Does anyone out there have a system for adding to a position that is going your way? Have you found something that works better for you than a Fibonacci sequence but still lets you build up a big position?
You are correct that the ES needs to move up 300 points without a 20 point retracement to make $17 million. But even a 120 point move results in a $13,000 profit. I'm trying to "cut my losses short and let my winners run" I think a system in this vein would achieve that goal. Attached is an Excel file of this simple system.
LOL thats the only thing that makes me money! I still feel I should make more money than I do. Probabilities in my favor (over 60%), have every winning trade with a full position, AND every losing trade is a small position. Still dont have a million bucks.
How does it test out in simulation? If you have the computer simulate a million parallel universes, each universe containing a clone of you who performs your investing strategy, how many of them achieve millions of dollars of profit in less than 5 years? If 900K out of 1M clones reach the goal, it indicates that your chances of reaching the goal, here in this one universe that we call "the real world", are 90%. On the other hand, if 50K out of 1M clones reach the goal, it suggests your chances are 5%. Might want to bifurcate your strategy, risk $900 per position rather than $1000, and invest the leftover $100 in a diversified portfolio of Lotto tickets. Oh yes, one more note: don't forget about gaps. The market can gap right past your stop and give you an unhappy fill. You can lose much more than your planned "risk per trade" because stops aren't always filled at your stated price. Darn it.