- Search Energy EXCH
-
algoTraderJo replied Mar 13, 2017Simple example of USDJPY Linear regression algo using 1H charts and 60 pip size renko bars: SL_ATR_MULTIPLIER=1.3 BARS_IN_FRONTIER=60 BARS_PER_EXAMPLE=2.0 NUM_OF_EXAMPLES=50 RENKO_SIZE_PIPS=60 image The ML algo here uses renko bar returns as ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 13, 2017You can even make something like this work with the 60 min timeframe if you use larger renko sizes. The renko function I posted is consistent, so whatever you get, even if there are some "errors" caused by the occasional OHLC bar that will be larger ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 13, 2017What I have done then is create systems that trade on the 5M in F4 and only trade when a new bar forms on the renko chart. These strategies use STRATEGY_TIMEFRAME=5 in the set file and passedtimeframe=5 on the tester configuration file, they also ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 12, 2017I also wanted to say that finding a consistent procedure to create renkos from OHLC was something I had to do since doing renkos from tick data was simply not feasible for me because of the computer power and time it requires. I am happy trading ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 12, 2017That's right indeed. Using a fixed point like zero as a fixed reference to build the renko grid can make your charts consistent but then you have the issue of picking that point and justifying why it is special. If you are making a 10 pip renko ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 11, 2017How do you make the renko charts immune to this problem? Any takers?
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 11, 2017The problem can be better understood when you think about how renko bars are made. For the sake of example let's say you want a 2 point renko chart and the market moves in the following way through 7 ticks: 1. 10 2. 11 3. 12 4. 11 5. 10 6. 11 7. 10 ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 11, 2017Renko bars can be complicated. If you have ever backtested renko bars using something like MT4 you'll notice that you never seem to get accurate results. Yes, even if you use tick data. You write a pretty EA, get very good results and then you see ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Mar 11, 2017At last I have some time! Having kids often means no time for fun on forums. Time to get back to where we were and make a few additional contributions here, for all those machine learning enthusiasts out there. I have been exploring the use of renko ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 21, 2016You can find an example of using pattern filters + hour filtering below. This is the code I used for the creation of the examples: RegressionDataset regression_i_simpleReturn_o_tradeOutcome_examplesByStructure(int period, int barsUsed, int ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 21, 2016Combining the pattern filters with the hour filters is actually what I was going to continue with. You can do this and you can get good results and one very interesting consequence is that you reduce the number of examples needed per retraining ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 21, 2016I am not using this thread to show you the "best possible system", I am using it to expose new ideas that can help you expand your number of potential ML systems. The pattern based inputs I showed are simply a quick demonstration of the viability of ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 21, 2016Measure the probability of over fitting at different amounts of data used using the method on the paper. For machine learning systems that always retrain (this is key) you will see that this value becomes smaller and smaller as you use more and more ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 14, 20161. Yes, I am referring to replacing the denominator with variance, which does gives Thorpe's continuous optimal f calculation of the Kelly criteria. 2. If the random walk is long enough you won't see any acf, especially if the negative drift is ...
Automated trading algo generation
-
algoTraderJo replied Jan 14, 2016I'm gonna show you a simple initial approach to the "pattern detection" idea. I will use the open of the last 7 bars to define a "pattern" which will be the relationships between these values. I then look into the past to see if there are enough ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 14, 2016No, Metatrader 4 is a terrible platform. I trade with Oanda, connecting directly through their broker API.
Automated trading algo generation
-
algoTraderJo replied Jan 6, 2016I want to float an idea that I now want to develop going forward on the thread. Up until this point we have built ML algos that have some characteristics in common: A system always uses the same number of examples through a test Examples are built ...
Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 6, 2016Coming back from the holidays I see you have made some very clever advances in ML at asirikuy. I will give your ideas a try, see if anything interesting comes up

Machine Learning with algoTraderJo
-
algoTraderJo replied Jan 6, 20161.1. I would be referring to the first, reducing the variance within single systems. I don't see a lot of advantage in reducing the variance from systems that come from the same process because the fact that a process generates a varied array of ...
Automated trading algo generation
-
algoTraderJo replied Jan 6, 2016Failure needs to come from some formal criterion, lines in the sand that are psychological in nature have many problems. A drawdown threshold that you simply come up with based on what you can tolerate might be a premature exit point because it ...
Automated trading algo generation