A system for riskless long-term profits in real markets

[sambian]

Quote from MTE:

There are at least two problems with your backtesting.

1. You don't take into account commissions and slippage. $800 difference between the two portfolios in the first example wouldn't even be enough to cover the commissions.

I'm just showing that the idea works when tested with real prices. This is how it's done. If you need examples of other people who didn't take into account comissions and slippage, but their work was appreciated, look at Black, Scholes, Merton etc.

Quote from MTE:

2. In reality you can't trade at the closing price, so the actual trade price would be different from the close. This may not sound like a huge thing, but it can add up to quite a bit over time. Essentially, your backtest is suffering from look-ahead bias. You assume you trade at the closing price yet you don't know what the closing price is until after the fact.

The moment in which the trades are done is irrelevant. As I said it could be day, week, month, year, whatever. It could be 1:46 PM. But I have historical daily and monthly closing prices for the examples. If you have historical data for the prices at 1:46 PM, send it to me and I will show you that the results will be very similar

 

[MTE]

Quote from sambian:

I'm just showing that the idea works when tested with real prices. This is how it's done. If you need examples of other people who didn't take into account comissions and slippage, but their work was appreciated, look at Black, Scholes, Merton etc.

A closing price is not a real price since you cannot trade at this price. Your portfolio returns $800 more than buy&hold over a 60-year period. This is not an outperformance when you take into account the costs associated with monthly rebalancing.
There have been a lot of academic work that shows theoretical outperformance, but at the end of the day, what matters is whether a system works in real life or not, and based on your results your system doesn't have any real life value. You can't compare the work of Black and Scholes (option pricing model) to your work. They provided a framework for pricing options and everybody knows that it doesn't work in the real world without adjustments. Your work, on the other hand, is a system that apparently outperforms the market.

Quote from sambian:
The moment in which the trades are done is irrelevant. As I said it could be day, week, month, year, whatever. It could be 1:46 PM. But I have historical daily and monthly closing prices for the examples. If you have historical data for the prices at 1:46 PM, send it to me and I will show you that the results will be very similar

It's not the moment in which the trade takes place that matters! What matters is that in your backtest you use the price that is not available until after the fact. Just google "look-ahead bias".

 

[sambian]

Quote from MTE:

based on your results your system doesn't have any real life value.

What do you think, would I post it online if it had big real life value? It has theorethical value, which you can not appreciate, and this is fine by me, I'm used to it.

Quote from MTE:

It's not the moment in which the trade takes place that matters! What matters is that in your backtest you use the price that is not available until after the fact. Just google "look-ahead bias".

As I already said, I use prices which are available to me. If you have better database, I will appreaciate it if you send i to me

 

[MTE]

Quote from sambian:

What do you think, would I post it online if it had big real life value? It has theorethical value, which you can not appreciate, and this is fine by me, I'm used to it.


As I already said, I use prices which are available to me. If you have better database, I will appreaciate it if you send i to me

I'm more into real life trading value, but if you are more into theoretical value then that's fine by me.

On a side note, this thread doesn't belong in the options forum, as it has nothing to do with options.

 

[sambian]

Quote from MTE:

I'm more into real life trading value, but if you are more into theoretical value then that's fine by me.

I'm also more into real life trading value, I just don't post my real life systems online

Quote from MTE:

On a side note, this thread doesn't belong in the options forum, as it has nothing to do with options.

MathAndLogic posted the other thread in this forum, so I decided that here is the proper place to post this thread, so that people can follow the discussion. And actually this thread has a lot to do with options, but I have to be very very theoretical to get to that and I'm worried that nobody will understand me.

 

[donahuedc]

I would be interested in how you feel this has to do with options. If you don't want to post it, I'd enjoy a private message.

 

[rew]

Quote from sambian:

1) Introduction

The system described in the article “A system for riskless long-term profits in efficient markets” was simplified as much as possible in order to be as understandable as possible. The examples were theoretical and the price movements were simulated using Excel. Here I present examples from the real world.

2) Important notes

2.1) The system is most profitable when the price moves up and down but at the end returns to its starting point. In this case the number of up movements is equal to the number of down movements, the price is the same, and our profit is maximum. The sequence of up and down movements is irrelevant – this is a property of the Kelly bet.

2.2) A property of my trading system is that it requires a long time to be profitable when the price moves away from its starting point. If the number of movements is large enough, the system will make profit – because of the “law of large numbers” we can expect enough up movements in the long run.

2.3) In the first article the theorethical examples were made by simulating price movements to a certain point, for example 100% increase or 50% decrease. But actually we don’t need to constantly watch the prices and be ready to readjust the portfolio whenever a certain point is reached. We have positive expected values also when we readjust the portfolio periodically, for example once a month or once a day.

2.4) The profits are “riskless” only in the sense that they are not riskier than the other possible strategies – for example “buy and hold”.

...

This is basically a variation of Alexander Green's "Gone Fishin'" strategy: http://www.amazon.com/Gone-Fishin-P...=sr_1_1?ie=UTF8&s=books&qid=1270663164&sr=1-1

You buy several distinct asset classes and rebalance at the end of each year.

Because you tend to sell more of whatever is overbought and buy more of whatever is undersold it may well do somewhat better than buy and hold over the long run. OTOH, because the decision of when to rebalance is very arbitrary I suspect that an active trading strategy should do better.

 

[olias]

Quote from sambian:

This is a huge topic and I haven't had luck to really discuss it with anybody who understands my point of view. But still I'll try to explain it:

I'm starting to get your drift. I thought you were crazy at first, but I do find this interesting now. Thanks. I haven't read all your posts, but I think I will investigate it and think about it some more

 

[trend2009]

As i said in another post, the profit only comes from the second order of taylor expansion. it does not worth the commission, let alone slippage.

Quote from sambian:

I'm also more into real life trading value, I just don't post my real life systems online


MathAndLogic posted the other thread in this forum, so I decided that here is the proper place to post this thread, so that people can follow the discussion. And actually this thread has a lot to do with options, but I have to be very very theoretical to get to that and I'm worried that nobody will understand me.

 

[sambian]

Quote from donahuedc:

I would be interested in how you feel this has to do with options. If you don't want to post it, I'd enjoy a private message.

I will surely write what it has to do with options, but I need more time, which I don't have these days.
Here is something briefly:
Are you acquainted with the Cox-Ross-Rubinstein option pricing model? Those guys claim to calculate "probability" of up movement and down movement, but in their view an up movement of x is equally likely as a down movement of x. So for example a price increase of 50% is equally likely as a decrease of 50%. This can be seen in their papers. And it is also in the schoolbooks, like the book of Hull for example.
Do you think that they are right? If not, what do you think is the probability of an up movement of 50% before a down movement of 50%? We assume that the price follows a random walk and the risk-free rate is 0. I think I know the answer and I know how to calculate the probabilities, but I'm not sure whether the way I do it is the best possible, so I'd like to see other peoples' thoughts on the subject.
Of course, knowing those probabilities can be helpful in trading options. But there are also other theoretical implications following from my system, for which I will write later.