Random Entry & Perception

[meanreversion]

This thread is disappearing up its own bottom, I'm not even sure what its point is any more.

To call it disjointed would be an understatement.

 

[BSD]

Thing is Toast only started this thread to rebut a point Rath and I made on another thread.

And to try and make his point truth went overboard.

This is the exact text where Van Tharp explains Bassos experiment:


"TRYING TO BEAT RANDOM ENTRY

I was doing a seminar with Tom Basso (see his sections in Chapters 3 and 5) in 1991. Tom was explaining that the most important part of his system was his exits and his position-sizing algorithms. As a result, one member of the audience remarked, “From what you are saying it sounds like you could make money consistently with a random entry as long as you have good exits and size your positions intelligently.”

Tom responded that he probably could. He promptly returned to his office and tested his own system of exits and position sizing with a “coin flip”-type entry. In other words, his system simulated trading four different markets and he was always in the market, either long or short, based upon a random signal. As soon as he got an exit signal, he’d reenter the market again based upon the random signal. Tom’s results showed that he made money consistently, even using $100 per contract for slippage and commissions.

We subsequently duplicated those results with more markets. I published them in one of my newsletters and gave several talks on them. Our system was very simple. We determined the volatility of the market by a lo-day exponential moving average of the average true range. Our initial stop was three times that volatility reading. Once entry occurred by a coin flip, the same three-times-volatility stop was trailed from the close. However, the stop could only move in our favor. Thus, the stop moved closer whenever the markets moved in our favor or whenever volatility shrank. We also used a 1 percent risk model for our position-sizing system, as described in Chapter 12.

That’s it! That’s all there was to the system-a random entry, plus a trailing stop that was three times the volatility, plus a 1 percent risk algorithm to size positions. We ran it on 10 markets. And it was always in each market, either long or short, depending upon a coin flip. It’s a good illustration of how simplicity works in system development.

Whenever you run a random entry system, you get different results. This system made money on 80 percent of the runs when it only traded one contract per futures market. It made money 100 percent of the time when a simple 1 percent risk money management system was added. That’s pretty impressive. The system had a reliability level of 38 percent, which is about average for a trend-following system."

http://books.google.de/books?id=_YM...AA#v=onepage&q=tom basso random entry&f=false

Nothing strange or untoward about that at all for any normal trader, standard, simple backtest of essentially a trend following system with a win rate of 38%.

And that is ALL the information available publicly.

But to get his point across the Toast starts lying and slandering a CTA with a great money management track record by publicly accusing him of having fiddled the data, even though the logical fallacy behind that accusation is mind numbing, why on earth should Basso have fiddled the data when he was merely trying to figure something out for himself, no normal person would lie to themselves when embarking on a fact finding mission !!

Toasts accusation:

As it is, there are 2 major flaws in this study :

1 - The approach taken to the study is flawed
2 - The test itself is NOT proof of random entry, rather they admit to curve fitting as well as omitting key data

Forgetting the additional logical fallacy about the study being flawed coz, hey guys, you know, it's flawed, part 2 is nothing but an outright lie he will never be able to back up, that they admitted anywhere to curve fitting and omitting key data !

In the day and age of Madoff accusing a CTA of such data fiddling is no joke is nothing short of serious slander.

 

[Shakone]

I am very fluent in mathematics. I don't believe an arbitrage argument like that works MrGecko.

If you wanted to prove something mathematically like this, you would have to make assumptions, for instance the market is a random walk. If you assume this, you can prove that no matter what strategy you choose, your expected value is the same as what you started with, less the transaction costs. That is proven and a well known result. If you believe in random walk (I don't) then you would quit trading altogether, because no money can be made from it except short term luck.

If you don't believe in random walk, then where do you start? It has X distribution? Is that reflective of the market on all time frames? Probably not. People often say that if we tossed the coin and took the entry with same stop and target then chances are 50-50. But there is no proof of this. I imagine it is often close to 50-50 at a lot of points, but taking into account some other info, there are plenty of points where it isn't, and the probabilities will be very different. We never know the probability at the time we take it or this would be an easy game. So trying to prove this mathematically is a non-starter unless you make false assumptions.

 

[meanreversion]

Can I ask what your mathematical qualification is, Shakone?

 

[Shakone]

PhD, meanreversion, specialised in financial maths area/probability

 

[meanreversion]

Nice! I'm definitely not arguing the (coin) toss with you.

There's a bunch of specious reasoning on this thread. It's very easy to say something that sounds plausible, e.g. "it's important to move your stop to breakeven as soon as possible, in order to retain profits and prevent losses". But is this actually borne out in real life?

Rather than assume this, or accept other commonly held beliefs, just go and backtest and see if there's any merit in it or not. My testing shows that random entry with coin flip (always in the market) allied with trailing stop, over sufficient markets, IS profitable. I've probably tested something similar to Basso and come up with similar results. Over a ten year period, there is never an instance of the random entry system losing money. That's good enough for me ---- exits are important, so focus on them as much, if not more, than the entry.

 

[MrGecko]

If you don't believe in random walk, then where do you start? It has X distribution? Is that reflective of the market on all time frames? Probably not. People often say that if we tossed the coin and took the entry with same stop and target then chances are 50-50. But there is no proof of this. I imagine it is often close to 50-50 at a lot of points, but taking into account some other info, there are plenty of points where it isn't, and the probabilities will be very different. We never know the probability at the time we take it or this would be an easy game. So trying to prove this mathematically is a non-starter unless you make false assumptions.

Imagine a market that could only go up, and you apply the heads/tails R:R = 1 strategy to it. You are still only going to get a win:loss ratio of 50:50 because you will be long 50% of the time and short 50% of the time. 100% of your long trades will be winners and 100% of your short trades will be losers.

Now change your r:r to 2:1. In a market that is only going up, you will still be long 50% of the time and short 50% of the time, except when you are long you make double what you lose.

My no arbitrage argument takes portfolio A and uses R:R=1, and hedges it with portfolio B of 2:1 R:R. The result of this two portfolio strategy would be profitable all the time, which is what you might expect in a market that could only go up.

Now introduce a down element to the market... Of all your Long positions, not all 100% will be profitable, but this is mirrored in your short positions which are no longer 100% losers. You could construct the same two portfolio's and expect the same results, over time you will always end up with a profit. That profit, however, cannot exceed that which the market offers, otherwise an arbitrage opportunity exists.

 

[BSD]

Nice! I'm definitely not arguing the (coin) toss with you.

There's a bunch of specious reasoning on this thread. It's very easy to say something that sounds plausible, e.g. "it's important to move your stop to breakeven as soon as possible, in order to retain profits and prevent losses". But is this actually borne out in real life?

Rather than assume this, or accept other commonly held beliefs, just go and backtest and see if there's any merit in it or not. My testing shows that random entry with coin flip (always in the market) allied with trailing stop, over sufficient markets, IS profitable. I've probably tested something similar to Basso and come up with similar results. Over a ten year period, there is never an instance of the random entry system losing money. That's good enough for me ---- exits are important, so focus on them as much, if not more, than the entry.

Great summary imho !



That apart this thread whose entire raison d'être was fataly flawed from the outset - Toasts unfounded allegation that Basso admitted to data fiddling - is really just going round in circles anyway.

Anyone who wants to be right more than make money, who demands a guaranteed profit irrespective of market conditions better go waiting for the green light from the great spaghetti monster in the sky !!



Shakone, good stuff, phd in maths, gotta respect that !

I just have a masters in politics lol.

 

[meanreversion]

I have a BSc in mathematics (first class!) but that is from 1994 and I've forgotten most of it now.

 

[meanreversion]

Imagine a market that could only go up, and you apply the heads/tails R:R = 1 strategy to it. You are still only going to get a win:loss ratio of 50:50 because you will be long 50% of the time and short 50% of the time. 100% of your long trades will be winners and 100% of your short trades will be losers.

Now change your r:r to 2:1. In a market that is only going up, you will still be long 50% of the time and short 50% of the time, except when you are long you make double what you lose.

My no arbitrage argument takes portfolio A and uses R:R=1, and hedges it with portfolio B of 2:1 R:R. The result of this two portfolio strategy would be profitable all the time, which is what you might expect in a market that could only go up.

Now introduce a down element to the market... Of all your Long positions, not all 100% will be profitable, but this is mirrored in your short positions which are no longer 100% losers. You could construct the same two portfolio's and expect the same results, over time you will always end up with a profit. That profit, however, cannot exceed that which the market offers, otherwise an arbitrage opportunity exists.

Sorry to say this, but this is what I meant by specious arguments.

What on earth is a market "that can only go up"? This makes no sense. If it can only go up, it will simply go instantly to infinity --- this is a nonsense proposition!

 

[MrGecko]

Sorry to say this, but this is what I meant by specious arguments.

LOL... no offence

My point is that over time you can't expect a random strategy to return anything more than the market rate, no matter how you chop and change your R:R.

Happy for Shakone to pull it to pieces.

 

[meanreversion]

At the risk of being a boring t*sser, you need to define "random strategy". Everything that's been outlined so far is a random ENTRY system.

A purely random system (ask the dog) is unlikely to be profitable, I would agree.

 

[MrGecko]

I mean a strategy that enters at random and has a pre-determined Stop/Target OCO set.

 

[meanreversion]

Ah right, I've tested that one as well and it isn't profitable in the long run, equity curve just fluctuates around its starting level (assuming zero slippage).

I've also tested entry using indicators (MACD, RSI etc) with 1:1 profit/loss ratio, or 2:1, or even 1:2, and that is not profitable in the long run either. That's a new thread in itself though, possibly.

 

[DionysusToast]

That might be true, but I have to say that you've made an awful job of getting your point across.

First you mentioned random system, which meanreversion pointed out doesn't make any sense, and not even relevant since we're talking about random entry only.

Then you make ludicrous statements like


My random entry system is to enter once only, and take profit at 10 pips, and my stop at 10 pips. CAN it make money? Of course. You are picking the wrong words to explain what you are trying to explain. There are no laws of mathematics that say you can't make money with a random entry, but there are laws of probability which say that you CAN. And if you can win money on one trade, you can win on 10 or 20 or 30 in a row.

So I can only guess that what you are actually saying is
1) The experiment in the text was flawed
2) That a random entry system will not make money in the long term (i.e. over a high number of trades) in all markets and in all conditions.

Well for 1 - - Agreed, it isn't even clearly explained

And for 2 - not every successful system makes money in all markets under all conditions. In fact, I'd guess that most successful systems don't in all markets - although I can't prove this.This is not really the fault of random entry. But you certainly haven't proven that with the right exit criteria and money management that it CAN'T be done. The reason you can't prove that, is because it is not provable. And as someone who claims to understand maths and physics, you should realise that there is simply no way you can prove that over ALL possible exit criteria and money management rules, a random entry system can't make money in the long term.

You would have to prove for EVERY (and there are infinite) exit criterion, you can't make money in the long term. Whereas those who say you can make money, only need one counter example, and you've already been given plenty.

If you are arguing something else, then try to explain exactly what you are arguing, otherwise you get a thread where people are arguing different loosely related topics, and as a result not agreeing.

Actually - your point 2 is pretty much my point.

not every successful system makes money in all markets under all conditions and that the Van Tharp system is just another system that makes money in certain markets under certain conditions.

As I have said - this system worked because of complimentary market conditions. The system outlined in the text THINKS it proves something about entering at random. It does not.

What it does do is to prove that a system that keeps re-entering the market with a trailing stop will make money over time in certain market conditions.

The 3 things that make this system profitable are :

1 - a fairly even distribution of heads & tails (hence only 80% working)
2 - a method of closing a trade if it is counter trend
3 - a trending market

Remove one of these and the system will fail.

Any directional system that exists purely based on mathematics, without any need for complimentary market conditions and without any inherent edge would be akin to perpetual motion.

In other words - if such a system were possible, it would be theoretically possible for everyone to have one and for them all to extract more money from the markets than they put in, which is impossible.

Let's say this thread wasn't invented. Let's say that I posted a thread about a moving average crossover system with the same trailing stop as the Van Tharp book describes. Wouldn't there be a furore over my claims that my non-random system would beat the market in all conditions ? Of course there would. This is because most people can relate to a moving average crossover system and understand the conditions under which it will fail.

If we frame this in terms that we are familiar i.e. a very common non-random system and then claim it will make money all the time because of a trailing stop - the lynch mob would stretch down the street.

The system was presented to show that money management can make money even when the entry is poor. It does not prove that at all. What it proves is that a trailing stop can make you money regardless of entry in certain market conditions.

 

[BSD]

Thing is Toast only started this thread to rebut a point Rath and I made on another thread.

And to try and make his point truth went overboard.

This is the exact text where Van Tharp explains Bassos experiment:


"TRYING TO BEAT RANDOM ENTRY

I was doing a seminar with Tom Basso (see his sections in Chapters 3 and 5) in 1991. Tom was explaining that the most important part of his system was his exits and his position-sizing algorithms. As a result, one member of the audience remarked, “From what you are saying it sounds like you could make money consistently with a random entry as long as you have good exits and size your positions intelligently.”

Tom responded that he probably could. He promptly returned to his office and tested his own system of exits and position sizing with a “coin flip”-type entry. In other words, his system simulated trading four different markets and he was always in the market, either long or short, based upon a random signal. As soon as he got an exit signal, he’d reenter the market again based upon the random signal. Tom’s results showed that he made money consistently, even using $100 per contract for slippage and commissions.

We subsequently duplicated those results with more markets. I published them in one of my newsletters and gave several talks on them. Our system was very simple. We determined the volatility of the market by a lo-day exponential moving average of the average true range. Our initial stop was three times that volatility reading. Once entry occurred by a coin flip, the same three-times-volatility stop was trailed from the close. However, the stop could only move in our favor. Thus, the stop moved closer whenever the markets moved in our favor or whenever volatility shrank. We also used a 1 percent risk model for our position-sizing system, as described in Chapter 12.

That’s it! That’s all there was to the system-a random entry, plus a trailing stop that was three times the volatility, plus a 1 percent risk algorithm to size positions. We ran it on 10 markets. And it was always in each market, either long or short, depending upon a coin flip. It’s a good illustration of how simplicity works in system development.

Whenever you run a random entry system, you get different results. This system made money on 80 percent of the runs when it only traded one contract per futures market. It made money 100 percent of the time when a simple 1 percent risk money management system was added. That’s pretty impressive. The system had a reliability level of 38 percent, which is about average for a trend-following system."

http://books.google.de/books?id=_YM...AA#v=onepage&q=tom basso random entry&f=false

Nothing strange or untoward about that at all for any normal trader, standard, simple backtest of essentially a trend following system with a win rate of 38%.

And that is ALL the information available publicly.

But to get his point across the Toast starts lying and slandering a CTA with a great money management track record by publicly accusing him of having fiddled the data, even though the logical fallacy behind that accusation is mind numbing, why on earth should Basso have fiddled the data when he was merely trying to figure something out for himself, no normal person would lie to themselves when embarking on a fact finding mission !!

Toasts accusation:

As it is, there are 2 major flaws in this study :

1 - The approach taken to the study is flawed
2 - The test itself is NOT proof of random entry, rather they admit to curve fitting as well as omitting key data

Forgetting the additional logical fallacy about the study being flawed coz, hey guys, you know, it's flawed, part 2 is nothing but an outright lie he will never be able to back up, that they admitted anywhere to curve fitting and omitting key data !

In the day and age of Madoff accusing a CTA of such data fiddling is no joke is nothing short of serious slander.

Buddy back up your slanderous lie already that per your phantasy running wild Basso admitted to curve fitting and omitting key data !!!

 

[DionysusToast]

At the risk of being a boring t*sser, you need to define "random strategy". Everything that's been outlined so far is a random ENTRY system.

A purely random system (ask the dog) is unlikely to be profitable, I would agree.

Actually - whatever the system is - it will be taking advantage of something that can disappear.

In the Van Tharp system, it takes advantage of trends to overcome the bad trades.

People that trade options can place trades on the volatility as opposed to the direction. You can 'bet' with an option that the underlying will become more volatile without knowing the direction.

In a similar manner, the Van Tharp system bets that trends will exist long enough to establish a trade which stays in for a long enough period to overcome the losses.

If trends exist for that long enough, then many random runs will yield results that are profitable because the underlying trends make it so. On markets that don't trend as well, the results will be bad.

So - just like an MA crossover with a 1:1 stop and target is a bet on direction, so is the Van Tharp system a bet on the markets trending in such a way that it returns positive results.

It is not a test of random entry - it is a test of a system that makes money from trends - albeit it in a rather roundabout way.

 

[DionysusToast]

Buddy back up your slanderous lie already that per your phantasy running wild Basso admitted to curve fitting and omitting key data !!!

Whenever you run a random entry system, you get different results. This system made money on 80 percent of the runs when it only traded one contract per futures market. It made money 100 percent of the time when a simple 1 percent risk money management system was added

The system made money 80% of the time over the data set.

Then they changed the system

Then the system made money 100% of the time over the same data set.

This is curve fitting.

- How many things did they try to get 80% to 100% ?
- Why didn't they test on different time periods/markets after changing the system ?
- Why no details on the size of the winning & losing runs ?
- Why no details on the impact of the 1% risk system on the data set ?

Anyone that works in the testing of similar systems (I have 22 years experience) will tell you that this is shoddy work and poorly described at that.

Note that Basso only ran his tests over 4 markets. The 10 markets in Van Tharps thread that we are discussing here were executed by someone else.

 

[meanreversion]

I detailed the results of my tests on an earlier post, which I ran today. However, just because a random entry system is always profitable over a ten year period doesn't mean a whole lot. It might lose for 3 years in a row before kicking into life. I can't imagine anyone being happy with using random entry and losing for 3 years on the spin, so it's never going to be applied by a real person. Doesn't mean it won't make money over time tho!

 

[montmorencyt2w]

Ah right, I've tested that one as well and it isn't profitable in the long run, equity curve just fluctuates around its starting level (assuming zero slippage).

I've also tested entry using indicators (MACD, RSI etc) with 1:1 profit/loss ratio, or 2:1, or even 1:2, and that is not profitable in the long run either. That's a new thread in itself though, possibly.

Doesn't your second paragraph tend to contradict BSD's statement/citation regarding that Commerzbank fund? (New thread would be good, provided BSD agreed to contribute to it).
I suppose one's point of view on this might depend on ones view of the applicability or not of back-testing to real-world live trading.