I've presented some of my trading strategies in this thread before, the last few with accompanying backtests going back as far as 25 years. It was mentioned, once or twice, that I would be giving my best trading strategies away. Well, this is just what I did. I offered them to the Bill & Melinda Gates Foundation. I found it to be the best outcome for my years of research.

I view this offering as my way to help people, more than I ever could alone. It is all explained in my latest paper: A Donor Within.

.

http://alphapowertrading.com/images/DEVX/ADonorWithin.pdf

The paper shows that it is by letting the Foundation's trust unit grow as much as it possibly can that the Foundation itself could do more. And that it might be in the pursuit of these portfolio management techniques that the Foundation could reach its goals.

The trading methods described in the paper could help anyone wishing to outperform over the long term. It covers trading strategies that have been presented here before. One of which was selected to do even better.

I think that any big portfolio could benefit from this trading methodology which simply says: accumulate shares for the long term and trade over the process. IMHO, there are many ways to do just that.

In my system the trend is everything what is not the white noise. The system is completely manageable so the trend is what I want that it is.

@Guy...Well posted. You have given me something to think about.

In my paper: A Donor Within, is explained how an existing trading strategy was modified to reach higher performance. The section: One More Thing, starting on page 30, explains the process.

The desired expectations were put on paper before any testing: increase position size by 10, and improve CAGR for the 30 stock portfolio over the past 25 years. Needed transformations were determined, and then the program modified and debugged on a single stock. It's only after the debugging that the strategy was put to the test on the whole portfolio, just once, and results recorded what ever they were (Fig. 16 in the paper).

Changing the objectives did not require that much program modifications; all done within a few hours. Scaling up was easy: multiply the position size and the initial capital by 10. The answer to that was already known. More elaborate program modifications would be required to increase the performance level. The program (1,860 lines of WL code) is governed by equation (6) in the paper: A(t) = 10*A(0) + Σ(10*H(1+g+T)^t.*ΔP).

To improve on performance, I wanted the reinvestment rate g to rise a bit as well as the contribution from the trading activity T. But, you can increase the reinvestment policy rate only if you generate more cash, and this has to come from adding more profitable trades. Therefore, the effort was put on increasing the trading activity over the entire holding period (25 years) which had for side effect to provide the needed cash to increase the reinvestment rate. No predictive powers at play, just the execution of what is implied in the equation itself.

Increasing the trading activity, within the same portfolio restraints, had for result to gradually pump new cash in the account which could then be reused depending on the reinvestment policies put into place. I transformed the whole payoff matrix as a single block, all 195,000 data elements even before making the portfolio test. There were no surprises once the test was executed. It did as asked to do by equation (6) from day 1.

There was no optimization done either, since none was needed or required. I would even add: very little could be done in the first place due to the nature of the strategy itself. There was only one portfolio test. All the program could do was follow equation (6). And even if you wanted to optimize something, the question would be: optimize what? Certainly not a technical indicator, none were used. The trading strategy is based on random-like entries. All I could do is say: do more of this or do less of that.

The strategy presented in the paper is not a conventional trading strategy. Instead of looking to optimize parameters or indicators, it executes mathematical rules and formulas. Its mission is to build a long term portfolio. To do so, it accumulates shares for the long term and trades over the process.

Equation (6) deals with the whole inventory matrix as a single block. It deals with a decision surrogate process which changes the inventory held from time to time. The buy matrix B has all its entries the result of random functions. Doing another back test would give totally different trades, but would still end up with close to the same answer due mainly to the sheer number of generated trades.

From Equation (6), it is also implied that one could take an ordinary trading strategy H(?) and transform it to comply with equation (6) thereby improving its performance. Each trading strategy has its own signature, but since the original version is: A(t) = A(0) + Σ(H(?).*ΔP), then equation (6) would transform that program to produce more. One, by scaling it up; and two, by finding ways to increase the position size with time.

There are many examples of such strategy transformations on my web site.

For those wishing to gain more insight into the trading methods I've presented, you can find a more elaborate description of my last post at:

.

http://alphapowertrading.com/index.php/papers/173-a-unique-approach

Hope it helps you in your own search.

In my last post I made reference to my last paper: A Donor Within. It dealt with a stock trading strategy designed to perform over the long term (some 25+ years). This new one is related to it, it's kind of a follow-on.

The program used was originally developed using Wealth-Lab software.

The strategy considered is the same: DEVX V6, a stock trading strategy designed to accumulate shares over the long term and trade over the accumulative process. It's primary mission is to build a long term portfolio. And it's singularity being that all trades are the result of random functions. It also shows why building a stock inventory over time has its own merits, and these merits can be viewed on the bottom line.

Since there are quite a few graphics, and it's 18 pages long, please follow the LINK:

.

http://alphapowertrading.com/index.php/papers/174-devx-v6-revisited

I think that it could be of help in designing your own trading strategies. Hopefully, you will find some interesting ideas or new avenues to explore.

Thank you for taking the time to read my research note.

Some other considerations not included in the last paper: DEVX V6 Revisited.

Since all the trading has for origin random entries and the results governed by mathematical functions, one could surmise that technical analysis was not part of the trading decision process, and I would have to concur. Just as fundamental data was not part of the solution either.

You have in the DEVX V6 enhanced stock trading strategy a peculiar status: no technicals, no fundamentals and yet, great performance. Not just for a few months or a few years on a single stock, but for the whole duration of the test: 25 years of data for each of the 30 stocks. Not the best stock selection one might make, but still an acceptable choice.

The payoff matrix: H.*ΔP is of size 6,586 rows by 30 columns. This is the same as having a big grid of 197,580 boxes over which are scattered over a million trades, over a million trading decisions, with the restriction that there is enough equity in the account to complete any of the trades. And since all the trades are randomly generated. Every test solution would be unique, no way to have twice the same solution no matter how hard one would try. On that, the odds are astronomical to say the least.

The trading strategy is not even an hybrid, it's none of the usual acceptable choices. Trading using mathematical formulas, using random entries, and yet the program executed better than just well.

Why should such a strategy win over the long term? I think that your ability to answer the question might help you in your own strategy design. What do you think?

There were no surprises in the revisited version. All behaved as expected. Making another test, I would again obtain an overall performance level close to the results already shown. I don't intend to do such a test for now, I already know the answer to that. But maybe later, say in a few months were there would be something new to learn. Or maybe if I have more improvements that I could bring to the code.

What's important in a trading strategy?

First, it has to pass a basic acid test. All trading strategies, what ever their origin, nature, or duration, can be expressed using the following payoff matrix equation: A(t) = A(0) + Σ(H.*ΔP). One has to design a strategy having Σ(H.*ΔP) > 0, meaning that a trading strategy must at least generate a penny in profit and this over the whole trading interval.

Sounds so basic and yet I often see strategies built to fail from the start and this by design. It's not a question of the idiosyncrasies of the market, it is that the trading strategy itself was really built to fail. Most often due to the author's inability to look at the whole picture.

That you trade once, a 100 times, a thousand or a million times it does not change the nature of the equation above. Did your payoff matrix exceed zero or not: Σ(H.*ΔP) >? 0. That is the question. It's not the size of the matrix that counts, it's the output. It is how much did your trading strategy produce over the entire trading interval under consideration. There is no need to play the game if after 20+ years your payoff matrix has a negative value: Σ(H.*ΔP) < 0. This would be a total waste of time and resources.

Each one of us can address the problem differently. I don't see a problem there. There is a multitude of solutions. Some think designing a moving average crossover system is sufficient to win, only to find out later that their portfolio had its own agenda with a special unscheduled meeting with oblivion. If the person had tested his/her strategy properly and thoroughly, it would have been evident that the future was not so rosy.

If a trading strategy could not survive its past, and over a sufficiently long trading interval; on what basis could one claim it might survive its future over the next 20+ years?

It is why I do long term back tests. At least, if my strategy could survive its past, and over a long haul, I think it has a much better chance than the one that did not.

I rarely see in this forum people presenting any long term summary performance reports generated by their trading software even if it is a byproduct of every simulation they make. But I do see chart setups showing hand picked examples that worked on short term past data.

Why is that you think?