Fruitless Pursuits ?

[counter_violent]

This is true, academics become very specialised, and experts in a very small area, which often has little application.

Perhaps the other message in the opening post is that both got what they were looking for. The academic convinced himself it couldn't be done, and so couldn't do it, and evven wrote a paper to justify himself. The wife just wanted the table in the room, and somehow found a solution. If you keep thinking that trading profitably can't be done, you'll find ways to convince yourself it can't (the market being a random walk seems to be a common example around here).

Convinced himself it couldn't be done using tools and methods known to him.

His wife on the other hand was a little more practical, she walked to the garage for a screw driver and took the legs off.

The moral of this story being...stick to the day job.

 

[DionysusToast]

In a bookshop today I read this anecdotal tale: Maths Professor & his wife at home need to move a large table from one room to another, through a narrow doorway. After much effort and re-positioning they give up, admitting it can't be done. Because he's that sort of bloke, Professor writes a mathematical proof confirming the situation. Meanwhile, when he's finished he finds that wifey has successfully trans-roomed the table!

Do you ever feel like that with trading?

There is an outstanding book by Edward De Bono called "I Am Right, You Are Wrong".

It explains how our brains work (or rather provides a tested hypothesis), how the human race currently resolves problems and the role of language in limiting our thinking.

I can see the effect of this clearly in my own thinking and on this site. For instance, people looking for a system with parameters that can be re-applied to make money and not looking for a more 'fuzzy' solution.

If you are searching for a solution to the trading problem, I think that it would be worth reading this book so that you don't effectively end up doing the same thing over & over again in your quest.

The way we think limits the kind of problems we are able to resolve, yet we are capable of changing the way we think. For instance predicting the weather and curing disease are good examples of areas that appear too complex for us to make strides in. The weather is deemed too complex because of all the variables. Doesn't trading also have millions of variables to consider too ?

Perhaps the efforts of attempting to take into account so many variables and analyse them is what hampers most people (especially engineers/mathematicians). Perhaps this explains the mythical "cockney barrow boy" trader who takes to it like a duck to water and gives you a blank look when you mention MACD....

 

[arabianights]

Arabian's impartial thread review:

shakone understands maths and provides cogent argument, while everyone else nit picks and misses the point. Free conspiracy thinking from shadowninja as well

(DT's post decent but fairly irrelevant)

 

[barjon]

................Well if you toss a coin and it comes up tails 3 times, it doesn't make it more likely to be heads next or more likely to be tails again. Because there is no information obtainable, because it is a 50-50 random walk. So no matter how it turns out during trade, you can't manage it so that it becomes profitable, nor does the risk reward ratio matter............

mmmm, my maths isn't strong but if you toss the coin 10,000 times you would expect around 5,000 heads and 5,000 tails or thereabouts. Thus, if your strategy is based on reversion to the 50/50 mean don't you have a positive expectation even though each individual toss remains 50/50?

jon

 

[0007]

There is an outstanding book by Edward De Bono called "I Am Right, You Are Wrong".

It explains how our brains work (or rather provides a tested hypothesis), how the human race currently resolves problems and the role of language in limiting our thinking.

I can see the effect of this clearly in my own thinking and on this site. For instance, people looking for a system with parameters that can be re-applied to make money and not looking for a more 'fuzzy' solution.

If you are searching for a solution to the trading problem, I think that it would be worth reading this book so that you don't effectively end up doing the same thing over & over again in your quest.

The way we think limits the kind of problems we are able to resolve, yet we are capable of changing the way we think. For instance predicting the weather and curing disease are good examples of areas that appear too complex for us to make strides in. The weather is deemed too complex because of all the variables. Doesn't trading also have millions of variables to consider too ?

Perhaps the efforts of attempting to take into account so many variables and analyse them is what hampers most people (especially engineers/mathematicians). Perhaps this explains the mythical "cockney barrow boy" trader who takes to it like a duck to water and gives you a blank look when you mention MACD....

Agree with that - especially "Perhaps the efforts of attempting to take into account so many variables and analyse them is what hampers".

All De Bono's books are good. Another excellent book if you're into this sort of thing, is Graham Polya's "How to Solve it" - uses an heuristic approach. I think this genre can be useful if you're tackling a project like Cable Sal's (http://www.trade2win.com/boards/trading-journals/100332-dont-panic.html).

For the curious, the book I was reading at #1 was Ian Stewart's "Hoard of Mathematical Treasures" : fascinating reading even if you don't understand maths but like puzzles. Whether or not it was Stewart's wife or table he doesn't say!

 

[DionysusToast]

*sigh* you really dont get it do you ?

the positive expectancy isnt obtained from the entry

This is merely your perspective.

You should realise that both you and the person you *sigh* at could both be right at the same time.

 

[DionysusToast]

Arabian's impartial thread review:

shakone understands maths and provides cogent argument, while everyone else nit picks and misses the point. Free conspiracy thinking from shadowninja as well

(DT's post decent but fairly irrelevant)

Allow me to explain.

Anyone looking at the problem may come to consider the adage "if the only tool you have is a hammer, you tend to see every problem as a nail". The professor stayed within his field of expertise whilst his wife resolved the issue..

Now - what if we found out the wife was a carpenter?

Of course, your perspective of the wife's wisdom in the issue would now change.In this case, we have 2 people that didn't stray from their area of expertise - the wife did the same as the professor, it's just that her skills were a 'fit'. Just a thought...

We could pat ourselves on the back for concluding that we need to use different skills to analyse and resolve an issue. That the problem here is we need to become proficient in multiple skillsets and apply them appropriately. Hence the need for more & more TA/Indicators/fundamental correlations in resolving the trading problem. In the end the one thing we don't attempt to change is the whole analysis process itself. So - before you go and cast the net wider, consider that you will go through the same thought processes with all the new info you get.

On the other hand, we could explicitly throw all that research away, go get a "cockney barrow boy" and ask him so show us why on earth he's driving around in a BRAND NEW Ferrari...

 

[MrGecko]

mmmm, my maths isn't strong but if you toss the coin 10,000 times you would expect around 5,000 heads and 5,000 tails or thereabouts. Thus, if your strategy is based on reversion to the 50/50 mean don't you have a positive expectation even though each individual toss remains 50/50?

jon

My hunch is that "no", you don't.

To start with, the deviation from the mean (say you had 1,000 heads in a row) isn't an input in the decision of the coin to land heads or tails - it's an indepenent random variable.

Secondly, and some of you may take this as a cop out, but you would need to know that the coin is fair in the first place, which is a pretty mighty barrier to this strategy working IMO.

I'm with Shakone. Around here I think most "retailers" have the Random walk stuff all srewed up. I have even seen people defend a broker who bestowed the virtues of trendlines plotted on randomly generated data in excel. And aren't there spread betting companies that will let you trade on a random index??

 

[MrGecko]

If you keep thinking that trading profitably can't be done, you'll find ways to convince yourself it can't (the market being a random walk seems to be a common example around here).

From my experiences of T2W the Random Walk Hypothesis (RWH) is both the average T2W'ers saviour and executioner. The common concencus around here IMO is that markets are RW's and that a positive expectancy strategy properly executed (along with money managements, discipline etc) is all you need to trade profitably, and so that answers all their "world view' questions and stops them cracking up - what is the saying, " God of the Gaps"? They put it all down to Random Walks and continue to tweak their strategy.

The downside is that the majority of their time (and money) is spent investigating trading methodologies that have little application to real world non-random walk markets, such as RSI crossovers on a 5m timeframe. Moreover, there is then an element of survivorship bias when someone makes some money using said strategy.

All IMHO and good trading to all.

 

[barjon]

My hunch is that "no", you don't.

............To start with, the deviation from the mean (say you had 1,000 heads in a row) isn't an input in the decision of the coin to land heads or tails - it's an independent random variable.............

MrG

I absolutely agree that the fact that you've had a run in one direction doesn't affect the odds in respect of the next single toss. However, if you say the odds on each individual coin toss are 50/50 then over a long series of tosses - the longer the better - the expectation is for an equal number of heads and tails to result.

Thus, your expectation is that any deviation from the 50/50 mean will be corrected in due course and you have moved from the probability attaching to a single coin toss to the probability attaching to a series of coin tosses. The bigger the deviation in one direction the greater the probability that there will be a series of tosses that deviate in the counter direction (ie: reversion to mean).

All assuming you've got an absolutely fair coin and spinner (the more appropriate word is frowned on and comes out ****** )

jon

 

[MrGecko]

MrG
... the expectation is for an equal number of heads and tails to result.

Thus, your expectation is that any deviation from the 50/50 mean will be corrected in due course and you have moved from the probability attaching to a single coin toss to the probability attaching to a series of coin tosses.

Hello mate, enjoying your Sunday? I am putting off doing the housekeeping

As for you point here, I'm afraid I disagree... if your expectation is that you get half heads half tails, then the expected deviation from where you are in the series is 0.

If you had a string of 1000 heads with a fair coin, all you can say is "from here on, I expect half heads half tails" , not "I expect to find 1000 more tails somewhere along the line to even things up" .

 

[DionysusToast]

Here's an offer to Barjon & anyone else that wants to take me up on it.

I will give 2-1 odds. Maximum bet $300.

We toss a fair coin with a fair person throwing (or suitable random number generation engine).

If the tosses come out even between heads and tails over 1000 tosses, I will pay you double your bet. If the amount of heads and tails is not 500 each, you pay me your bet.

Does anyone want to take me up on it? Note that each player will have to go through a seperate 1000 tosses, not all players on the same 1000 set.

Also - does anyone want to offer better odds than I have put here? If so - how high will you go?

 

[DionysusToast]

Why not 4DL? Surely with 1000 tosses, probability favours the outcome to be even.

Should we up the number of tosses to make it fairer? Maybe 10,000 tosses or 100,000 tosses.

If not - what odds would be fair at 1000 tosses? I think doubling your money is a sweet deal.

 

[bangkoker]

I like the idea 'toast but even if it deviates to 50/50 in the long run it inevitably will 'wander around' straying off the 50/50 at points, so to be spot on 50/50 at a set point in time is not likely..


Mind if you upped the numbers presumably it would work, I mean a mililon+, excluding times when the coin lands on the rim ofcourse

 

[Shakone]

mmmm, my maths isn't strong but if you toss the coin 10,000 times you would expect around 5,000 heads and 5,000 tails or thereabouts. Thus, if your strategy is based on reversion to the 50/50 mean don't you have a positive expectation even though each individual toss remains 50/50?

jon

No you don't. Lets say there have been 500 tosses, 400 heads 100 tails. So heads 80% of the time. Now if you keep tossing the coin 10,000 times, the % will move from 80% to something closer to 50%, but that is not as a consequence of there being more tails than heads. Lets say in the next 1,000 tosses you get 600 heads and 400 tails. Then you now have 1000 heads 500 tails, or 66.6% heads. It is moving closer to the 50%, but there weren't more tails than heads in that last ,000, there were less.

Actually, if you are tossing a coin 10 times, and it has come up 10 heads, then the logical decision is to bet heads again, and NOT tails. Because the fact that it has come up heads 10 times, doesn't make it more likely to be tails next, but might indicate the coin is not even fair, and is more likely to be heads.

 

[MrGecko]

the probability of this is happening 0.025 + change.

to be at 0, we need # wins to equal # losses. Over 1000 trades, this is 500 wins and 500 losses.

Because the trade results are boolean-ish (every non-win is a loss), we can do:

Number of combinations is:

1000C500 = 270E297

and the probability of each is:

0.5 ^ 1000 = 9.33E-302

=> Probability of breaking even is

270E297 * 9.33E-302 = 25.225E-003

 

[DionysusToast]

the probability of this is happening 0.025 + change.

BOOOOOOOOOOM

Yup - roughly 1 in 39 chance of you getting 500 heads and 500 tails on 1000 tosses.

If you increase the number of tosses - say to 1 million, you do not make the chances of an equal number of tosses more likely.

So - obviously my 2-1 odds are pretty ****ty when you consider the 39-1 chance of winning. Thanks for killing my new business venture Mr G.