Thu 3 Aug 2006
Well, on Monday I started trading the Uberman’s Portfolio with real money.? The yield is 10.99% at the current leverage.? I made this move after a month of forward trading (24/7 exposure)?and then completing 3 years of backtesting.? They both confirmed the theory that the porfolio does indeed stay tied to the “index” of the interest rate yield.? Random indexes of currencies did not do this.? This has given me the confidence (and data for position sizing) I need to dip my toe in the water.
Hopefully, this will be exactly what I meant about having an approach that allows you to benchmark your performance.?? By knowing that the portfolio is mathmatically driven to “stay in the lines” as best as possible and the road is mapped out ahead of time (the yield), then I have more confidence about the results and what they mean.? If after a year the return is close to the yield and walked the line the whole way, I can be reasonably sure?that the results were not luck but a result of the model at work.? I know that theory by confirmation isn’t the best form of proof, but in trading that may be all we get.
August 3rd, 2006 at 3:46 pm
Are you starting with the 11% yield based on your drawdown tolerance during your test period? I remember you saying something about the best possible allocation and it yielding north of 25%… Maybe I heard something incorrectly…
August 3rd, 2006 at 4:30 pm
Yes, it is based on my drawdown tolerance. I’ve lowered the leverage about in half just to be on the safe side. The current level I’ve chosen also now incorporates backtested data so that it’s more accurate in terms of position sizing. Before backtesting, all I had to work with is expected volatility (basically a measure of current volatility) because I had no actual trades to show me what the swings of the past were like. But that said, with compounding I expect a higher “return” than the simple yield. I’m happy with 10.99% right now. Perhaps later on I’ll feel more comfortable raising my tolerance level and thus my return. I might be playing it too safe but as long as I’m beating the risk-free rate by a resonable amount, I can afford to ease up slowly.
August 5th, 2006 at 1:58 pm
I like your cautious approach. The mechanics of your portfolio should be fairly consistent with a less leveraged position–good call, and a cheaper way to test out your ideas while still using real money. Could you refresh my memory and remind us of your risk tolerance (both the amount of draw-down and the period of acceptable draw-down)? Also, what measures beside’s lowering your leverage or cashing out your position would you take as a response to draw-downs approaching your tolerance level?
August 5th, 2006 at 8:14 pm
I am working with an average expected maximum drawdown (lots of adjectives there) of 10% but that doesn’t mean that 10% will never be exceeded. It simply means that that’s the middle of a confidence interval. Anything is possible and that’s why you work with smaller leverage. If everything went down at the same time then the correlations would quickly become such that the system might flash an “optimal” portfolio that couldn’t be expected to beat the risk-free rate, so at that point you would move to cash.
You can always reach ruin from any point. The only reason you manage risk is to make it harder to reach that point, not eliminate it all together. Even the good old stop loss is no guarantee because you might have many hit in a row, adding up to ruin. Leverage and cashing out are my only tools. If a six-sigma event occurs, so be it. I tried. There should always be some money you don’t ever risk and, technically, that is a form of leverage control. If I put $10K in a trading account and $10K somewhere else, I’m effectively using 0.5:1 leverage of my wealth. At the same time, however, if you are building that very wealth you don’t plan to risk later, you may have to risk it now.? But the everlasting question is how do you ever know if you are taking too much risk, or not enough, in order to reach your goals (e.g. a certain net worth by a certain time)? Maybe a good topic…