Mathematical trading

[tomorton]

............... the old technical and fundamental analysis methods are deemed to fail...........

On the contrary, these methods are proven to succeed. They only fail if mis-applied, as they usually are by private individuals.

However, in the absence of any other information from you as to your methdodology, I have to think you're using TA. This is a mathematical method, clearly pointing towards price action not fundamentals. It surely cannot be the numbers themselves which are important, as these are just abstract statements of a relationship between base currency and counter currency. So it must be the change of those numbers over time that gives you the cue to enter a trade. That's what TA is: this is just TA without looking at the chart, its just a chart expressed as numbers/formulae.

If I'm barking up the wrong tree, its only through lack of information as to your methodology, and I'd be happy for that to be corrected.

 

[t.roland]

On the contrary, these methods are proven to succeed. They only fail if mis-applied, as they usually are by private individuals.

However, in the absence of any other information from you as to your methdodology, I have to think you're using TA. This is a mathematical method, clearly pointing towards price action not fundamentals. It surely cannot be the numbers themselves which are important, as these are just abstract statements of a relationship between base currency and counter currency. So it must be the change of those numbers over time that gives you the cue to enter a trade. That's what TA is: this is just TA without looking at the chart, its just a chart expressed as numbers/formulae.

If I'm barking up the wrong tree, its only through lack of information as to your methodology, and I'd be happy for that to be corrected.

It's fine if you believe that. But I can't discuss the inner workings of my method and how it has been designed.

 

[timsk]

It's fine if you believe that. But I can't discuss the inner workings of my method and how it has been designed.

Hi t.roland,
Thank you for your reply to my question earlier in the thread.

Given your comment to Tom (quoted), are you going to provide your EA for free to enable members to load it into their MT4 platforms and take identical trades to you, or do they have to stay permanently logged in to T2W waiting for the bi-annual tip-off to get long or short EUR/USD? Personally, I hope it's the former but, with my official T2W hat on, I hope it's the latter!


Tim.

 

[t.roland]

Hi t.roland,
Thank you for your reply to my question earlier in the thread.

Given your comment to Tom (quoted), are you going to provide your EA for free to enable members to load it into their MT4 platforms and take identical trades to you, or do they have to stay permanently logged in to T2W waiting for the bi-annual tip-off to get long or short EUR/USD? Personally, I hope it's the former but, with my official T2W hat on, I hope it's the latter!


Tim.

I will not give my method for free and for money neither, but I am more than happy to share the trades I take, here, on your forum. And I hope that 2016 will bring more trades than before. There were already two, but usually there are more, and only long EUR/USD

 

[forker]

To be fair, investment banks do use mathematical models to price test. I know this to be true because I have worked an several of them in the past. They use models such as volsurface, interest rate sabre, and time series to name a few. They also use complex models for exotics where for example they extrapolate price points (tenors) and interpolate theoretical tenors. All investment bank trading is done off the back of models to manage risk and track valuations. All this being said I know that it's possible to make money without using these which I have been doing for years. I wouldn't say mathematics is any better than other methods the same as I would say that other methods aren't any better than price models.

 

[counter_violent]

I will not give my method for free and for money neither, but I am more than happy to share the trades I take, here, on your forum. And I hope that 2016 will bring more trades than before. There were already two, but usually there are more, and only long EUR/USD

Now it isn't a mathematical method if it's only long EURUSD. The back test results show long and short trades ! So what's going on?

 

[EnlightenedJoe]

I will not give my method for free and for money neither

Personally I don't care about the method so long as it works. But I have a gut feeling your mathematics is waiting for 20 week MA cross over.

 

[t.roland]

Now it isn't a mathematical method if it's only long EURUSD. The back test results show long and short trades ! So what's going on?

Where exactly can you see short trades in the statements and in the backtests?

Regarding if it's mathematics or not, I said what I had to say. I will post the trades here as they happen, in real time. Whoever wants to test them, should test them. Who doesn't, will see the final results anyway, because I will be posting that too.

For short, there's a different pattern, it's logical for it to be that way, I'm still working on it since it's not finished.
For short I only have a H4 which isn't as good as the long. I'll be posting a backtest soon.

 

[counter_violent]

Where exactly can you see short trades in the statements and in the backtests?

Regarding if it's mathematics or not, I said what I had to say. I will post the trades here as they happen, in real time. Whoever wants to test them, should test them. Who doesn't, will see the final results anyway, because I will be posting that too.

For short, there's a different pattern, it's logical for it to be that way, I'm still working on it since it's not finished.
For short I only have a H4 which isn't as good as the long. I'll be posting a backtest soon.

Oops, my mistake, was looking at wrong figure.

 

[forker]

Is there a model you are using or is this based off your own research?

 

[t.roland]

Now it isn't a mathematical method if it's only long EURUSD. The back test results show long and short trades ! So what's going on?

This is the short H4 which I don't use. I'll keep working on it to make it better.

 

[tomorton]

It's fine if you believe that. But I can't discuss the inner workings of my method and how it has been designed.

I was trying to provoke you into disclosing a bit about how you're doing what you're doing. Then both of us might learn as a result of pooling mental resources and experience. Otherwise, why come onto a public trading forum?

 

[t.roland]

I was trying to provoke you into disclosing a bit about how you're doing what you're doing. Then both of us might learn as a result of pooling mental resources and experience. Otherwise, why come onto a public trading forum?

I agree to post my trades here but I won't provide details about the method. One reason for this would be the amounts of money spent when I developed the robot. For example, I couldn't create it by using the default MT4 historical data because it's inaccurate. I needed tick by tick data, so I bought it, starting from 1999.
Only this data package alone costed me 25k euro, because I needed it to make the robot accurate.
I'm sure you understand this matter.

 

[t.roland]

Hello. on EUR/USD I don't have a pattern yet, but lately I have been testing my mathematical methods with success on stocks.

I opened a small trading account and I will post my trade entries here for the people who trade stocks.
Likewise the EUR/USD method, the stocks method has a success rate that is close to 100%.

Diversifying a portfolio is always a good idea, so I chose to try stocks as well.


All the best.

 

[t.roland]

Hello,

I also found a mathematical pattern for shorts.
Today I took my first trade.
I invite you to test this trade, its mathematical and the backtests show a 92% success rate. For now I do not have a fixed take profit, but I will update you once I close the trade.

Because it's the first trade taken on a live account, I chose to risk only 1% of it. I recommend you do the same, but no more than 3%.

 

[kalott]

If you claim that mathematics can predict the behaviour of banks who operate the financial markets then the mathematics must be lying. The banks are run by people. So precisely what they will do is unknowable. There are no known branches of mathematics that can predict human behaviour with any amount of certainty.

What you have got yourself into there is a typical beginner's trap. You curved fitted the past data and believe it has predictive capabilities. All beginners start that way.

Markets exist for the profit of their operators. That means the market rarely go to places where their operators make a loss. Sometimes they do, but these are very rare exceptions. The rationale is very simple: what idiot would put in all the money and effort to run a market only to make a loss ? Banks have tall shiny buildings precisely because they are very profitable.

It's understandable you are feeling euphoric that the mathematics have made you a profitable first trade. But if you are blinded by that, you are going to say bye bye to your 150k pretty soon.

Mathematics is particularly bad for predicting the markets precisely because it doesn't change. Markets are operated by people motivated by profit. There is no reason why their behaviour should be magically controlled by your mechanical mathematics. If a profit can be made by not behaving mechanically, that is exactly what they will do. People will do anything to turn a profit, including defeating mathematics.

Mathematics is better suited to predict the roulette ball because it is mechanical and has no motivations. I reckon you will do better at the casino if your mathematics is really that good. Even casinos are run by people. So your mathematics will lose against the ball tampering. See, the human motivation is super natural in essence ?

http://www.wsj.com/articles/jury-gi...oldman-programmer-sergey-aleynikov-1430496291

http://www.valuewalk.com/2016/04/goldman-sachs-group-inc-gs-lost-decade/

 

[kalott]

Hello,

I also found a mathematical pattern for shorts.
Today I took my first trade.
I invite you to test this trade, its mathematical and the backtests show a 92% success rate. For now I do not have a fixed take profit, but I will update you once I close the trade.

Because it's the first trade taken on a live account, I chose to risk only 1% of it. I recommend you do the same, but no more than 3%.

Interesting, that would be a long DAX at the same time if this play out?

 

[t.roland]

Interesting, that would be a long DAX at the same time if this play out?

I am testing now several indices: DJ30, NAS100, SP500, DE30 and GB100.

As for DE30, on backtests the bullish pattern appeared in 17 February and there are no signs that the trend is weakening now.

I will post here as soon as I will have any trades available for indices.

H4 and Daily are my current focus timeframes.

On EUR/USD, I'm looking at a target between 1.1015 and 1.0920.

 

[IFeelFree]

...that the old technical and fundamental analysis methods are deemed to fail.

I agree that, from a statistical viewpoint, market prices are very nearly random walks, and so forecasting should not be possible, or at best, very poor. Yet some people make money trading. So there's more to it than what statistics reveals.

The few that make profit in this market, use mathematics.

I agree that one can use mathematics to trade profitably, but there are lots of other people who trade without using any sophisticated math and also trade very profitably.

Technical and fundamnental analysis, which are based on the psychology of people and the market participants, do not work properly because psychology is constantly changing, but mathematics was the same 30 years ago and it will be the same in the future too. Some things do not change.

I would say that mathematical algorithms that are dynamic and adaptive can deal with changing market conditions based on shifting trader psychology.

...my motivation as a mathematician and statistician is rather a scientific one, to prove and show that mathematics in trading is net superiour to any other type of analysis. Mathematics does not lie.

Would you care describe what kind of mathematics you use, without giving away anything proprietary, of course? For example, I've developed software to implement statistical arbitrage, namely a cointegrated portfolio of multiple leveraged ETFs. I've coupled this algorithm with a vector autoregressive model. I trade once a week and after 8 months, I'm on track for >50% yearly return. Backtesting over 6 years shows less than 10% maximum drawdown.

 

[t.roland]

Congratulations for a excellent profit, too few peopleknow what it means to do real profit in trading, all are running for 200%/month and get to break their necks.

Without going into too much detail, I use PhD level complex mathematics which includes between other things, time series analysis.

But in the latest method of trading that I found and which I'm still working on and which gave the latest EUR/USD short trade, I also added geometry.