Wealth(t) = 1.0 for trading, 1.0 "reserve" => 2.0 first bad-trade Wealth(t+1) = 0.5 + 1.0 => 1.5 => 0.75 for trading, 0.75 "reserve" second bad-trade Wealth(t+2) = 0.75 + (0.5)(0.75) => 1.125 => 0.5625 for trading, 0.5625 "reserve" NOTE: You are now at a nearly 50% drawdown after only 2 bad consecutive trades Wealth(t+3) = 0.5625 + (0.5)(0.5625) = 0.84375 NOTE: You are now at ~60% drawdown after only 3 bad consecutive trades Wealth(t+4) = 0.421875 + (0.5)(0.421875) = 0.6328125 NOTE: You are now at ~70% drawdown after only 4 bad consecutive trades. In my books, that's a blowout. This strategy is not unlike Martingale, as has been pointed out. The only way to win in the long run is to have an infinitely large bankroll.
It is an excellent book and I would recommend it to anyone. His set-up was 50% stock, 50% cash and rebalancing after moves up or down. The book claims that this system is now known as "constant-proportion rebalanced portfolio". The problems with this system noted in the book are commissions and that no stock is as volatile as the examples of doubling and halving over and over again. It works, you just need something volatile enough (and best if it reverts to the mean). As pointed out in the book, if you found a stock that doubled or halved every day, you could turn $1 into a million dollars in about 240 trades (similar to your example). Joe.
Price does not have to get back to the starting point to make money, you proved that with your gold/silver example. As long as the small movements that head back towards the mean are greater in aggregate than the absolute movement from start to finish, your method makes money.
Cool. This means that I have reinvented the wheel So actually all the criticisms that I got here should have been directed at Claude Shannon
Actually, sambian, regardless of whether you're happy to answer my question (I hope you are), I have to apologize to you and admit that I initially behaved like a famous idiot. I dismissed your idea (which is still erroneous) for all the wrong reasons. The subject that you have stirred up is actually pretty deep and has lots of interesting connections (among them, as pointed out by others, CPPI, information theory and universal portfolios, which is where Shannon comes in). I think this is actually the most relevant/recent paper on the subject, but I can't find the PDF for it: http://repository.upenn.edu/dissertations/AAI3179708/ Now we can return to the discussion of our game...
I find it interesting that although I think the initial example was b.s. this thread is starting to sound like it will lead to volatility pumping. Which is something which has serious merit.
Apologies if this has already been suggested... Sambian: Would this work with going long on both an ETF & its Inverse twin at the same time and averaging daily? Anyone care to backtest it?
Sort of along these lines, an old post about "straddle scalping" QQQ. The long straddle protects against a large loss, but the sum of the scalps must be more than theta and friction losses: Option_Attack Registered: Sep 2002 Posts: 146 03-21-03 07:18 PM I've tried this in slightly different forms, but never was able to do more than b/e. A big part of my problem was that I was paying higher commissions and another was that I tried to do it on a EOD basis. Intraday should be much more profitable. My method was more "seat of the pants", than a strict "delta balancing" method. But I did examine my positions with a options graphing program. To my way of thinking, the options are mostly for protection and to provide a point of reference. The money is made by the stock scalps. Here is an example of what I tried on QQQ. QQQ is probably not the best stock to do it on, but the 1pt strikes and liquidity are nice. You must have cheap commish, or you will get killed. The basic theory is that most of the time QQQ will gyrate around where it is now, with many scalpable moves but in unknown directions. This method does not care which way it goes - we just follow it around. There are plenty of ways to try this. You can use a short straddle to help pay for a long strangle (10 short straddles and 20 long strangles), or you can just use a long straddle. To demonstrate, this example uses a simple long straddle. Assuming you tried this Monday with QQQ at 27.17 now: Buy 10 Apr 27 Calls for 1.3 Buy 10 Apr 27 Puts for 1.1 The cost is $2400. Theta is about $300 a week for the first two weeks and gets a lot worse after that. We have to at least make this back with stock scalping or we will lose our butt if QQQ doesn't move much. We will scalp up to an 800 share position. We have 1000 shares deltas from the long strangle, but we want to make sure our stock is covered from loss as we will be adding to 200 share positions, so we are only going to go up to 800 shares. Ficticious Trades: Monday - QQQ opens at 27 and moves up 0.25. We Short 200 shares. This effectively "locks" in gains "using" 2 of the Calls. It drops back to flat and we buy back the short shares, banking $50. There is some bad news and it drops to 26.75. We Buy 200 shares to lock gains for the puts. It drops down to 26.5 and we buy another 200 shares. We aren't worried that it will crash from here because that would make quite a bit as we are only long 400 shares. Short closing drives QQQ back up to 26.75 near the close and we sell back the 200 shares we bought at 26.5 and bank another $50. We are still long 200 shares from 26.75. Being long or short overnight has no meaning, as any major move will be a profit making event from the long straddle. Tuesday - Saddam rumors pop QQQ up to 27.5 before the open and we sell our 200 shares for a $150 profit. We then Short 400 shares because we moved up 0.50 from our "basis" at 27. QQQ drops before lunch down to 27.25 and we buy back 200 shares for another $50 profit. QQQ closes back up at 27.5. We are still short 200 shares from 27.5. (edit: You could have closed all 400 short here at 27.25 - no hard and fast rules) Wednesday - Quiet day and we don't get enough movement for a scalp. It sux, but it happens. Meanwhile, theta is stealing our money, but we made $300 from scalps so far. Okay, this should demonstrate how I played it. What should you do if QQQ kept moving up after you shorted 200 shares, then another 200, then another 200? Well, at some point you will have to roll the straddle and "reset" the whole game. Depending on how far and how fast QQQ moved and how many scalps you got, you might not have a profit when you roll the options. Well, of course it isn't quite that easy but that's about what I did and like I said, I wasn't able to do much more than b/e. How would this work long term? No idea.
Thanks for the info on gamma scalping. Your result would seem to shadow other people's attemps. I read somewhere (maybe here) that scalping gamma is more lucrative for premium sellers -not buyers- as time decay is in their favour. Still the question about ETF/invETF remains.