Thanks for the input. This is the result using the Euclidian distance of the 39 past bars returns. The "prediction" of the past is pretty bad but the variance seems lower too. image Now if I average both metrics: image How important is the ...

Machine Learning with algoTraderJo

How many bars is the best is part of the machine learning scope.

Machine Learning with algoTraderJo

Here is a very good and comprehensive introduction to machine learning. url ( url )

Machine Learning with algoTraderJo

This is exactely the Gambler's Fallacy: out of 100 trades there is absolutely no reason that there will be 50 winners and 50 losers. No I didn't try to enter more than 10 trades because it is a pain in the... neck to enter the values. It if was ...

Money management - What is the secret?

I just posted the result of your own software... image image

Money management - What is the secret?

I'm amazed by your results! Each time I try to cluster FX data I get a single cluster that encompass the whole dataset. That's why I tried another approach. Instead of clustering then testing in which cluster the query vector is I select the N ...

Machine Learning with algoTraderJo

The results are fully in line with my previous posts. With a winrate=50% and RR=1 your MM starts betting 7.37% of the starting capital that I set to 1000. If we bet 7.37% each time and get 10 losers the final result is 464.96. With your MM the ...

Money management - What is the secret?

Can you post a user's manual, please. The Help menu doesn't work. Even after google translation it is quite confusing. What are "quote iniziali uguali" (initial shares equal) and "quote iniziali diverse" (several initial shares??)? Why shall I enter ...

Money management - What is the secret?

Too bad the mode of the histogram will be exactly on the current price

Machine Learning with algoTraderJo

An histogram (empirical probabilities) of the move of the price from the current price. So I can get a target, a stop, a probability of this move. This give the direction, the TP, the SL and the risk to put on this trade (Kelly = expectancy / RR)

Machine Learning with algoTraderJo

I never tried to predict the state beyond a polynomial extrapolation which has absolutely no predictive power. The future trending state will be the same as the current one plus a random drift: linearly increasing variance. The trend is the red line ...

Kalman filter and currencies strength

The idea was to select a basket of three "probably trending pairs" by pairing the three strongest and the three weakest currencies. The pairing is made by taking care of their correlations. Using this method I get weights to apply to the MM: they ...

Kalman filter and currencies strength

Subscribed too.

Machine Learning with algoTraderJo

hmm again as you don't give the algorithm I can't be certain of the following but... you increase the size when your recent history winrate is below the average winrate and you reduce the size when the recent winrate is above average. When you get a ...

Money management - What is the secret?

What if you get a very longer than usual string of losers?

Money management - What is the secret?

You don't give the algorithm to know the next bet. I may be wrong but from the examples you give it seems to me that you expect some sort of mean reversion of your win/loss sequence. When you win "too often" according to your stats you bet less. ...

Money management - What is the secret?

Perhaps better if I post a chart of G/A as well with the oscillator image

Kalman filter and currencies strength

I actually stopped using KF for this because the lag is too high. I changed for something which is also more simple (but still lags!). To know the strength of the daily I use hourly candles (oversampling). To get rid of the unit of the prices to ...

Kalman filter and currencies strength

I completely agree with this.

Extracted Thread (per thread starter's request)

Yes sure but that's wasn't my point. I say that two periods of the same length may have totally different outcome.

Extracted Thread (per thread starter's request)