- Search Crypto Craft
-
FXEZ replied Aug 30, 2015I've done something like what Atokys suggested using the zig-zag indicator as an "optimal" representation of the trend or trade-able swings. This isn't too different from what you described earlier, perhaps just a different way of visualizing ...
Quant basket trader? I need your help!
-
FXEZ replied Aug 16, 2015beachboy44 and KoBo, If you create the categories manually like this, you implicitly assume that you will get more diversity and less related systems, an assumption which may be true or false. In reality, how do we know in advance which systems in ...
Machine Learning with algoTraderJo
-
FXEZ replied Aug 16, 2015University of Oxford Machine Learning course.
Machine Learning with algoTraderJo
-
FXEZ replied Aug 15, 2015Using a correlation filter to pre-screen (eliminate) systems before combination makes a lot of sense in terms of building an ensemble. One of the strengths of ensembles is the ability to combine diverse signals in a synergistic way. If the system ...
Machine Learning with algoTraderJo
-
FXEZ replied Aug 15, 2015KaBo, from the abstract of the paper I referenced in my last post: If you wish to do something seemingly modest like choose the best 100 systems to combine out of a list of 200 systems, the problem quickly becomes intractable as there are ...
Machine Learning with algoTraderJo
-
FXEZ replied Aug 14, 2015I'm very interested in learning more about Ensembles so algoTraderJo, appreciate the topic coming back to this area! On this subject, I saw a working paper that addresses Backtesting strategies based on multiple signals (also attached). The author, ...
Machine Learning with algoTraderJo
-
FXEZ replied Aug 9, 2015ezcurrency, since you brought it up here, my current thinking regarding why cointegration has trouble when applied to currency pairs has to do with the linearity assumption of regression. I think it's too much to ask two pairs to move in the same ...
Quant basket trader? I need your help!
-
FXEZ replied Aug 7, 2015And it isn't even that simple. The mean can drift so those 3 positions at the 3SD level can tun into a loss at mean reversion. This is why I don't put much faith in mathematical models that rely on a zero mean for doing such computations. I think ...
Quant basket trader? I need your help!
-
FXEZ replied Aug 7, 2015PipMeUp, If the market you are targeting is not really mean reverting, then the question "find the optimal..." is ill posed. If the market is really mean reverting then any simple strategy to maximize the number of trades * (open - close) prices ...
Quant basket trader? I need your help!
-
FXEZ replied Jul 24, 2015You can run FXTrade on multiple computers for the same account. One computer's FXTrade may disconnect, but the other computer may stay connected, even with the same internet connection on both computers. Use that as a way to diversify your ...
OANDA Discussion
-
FXEZ replied Jul 22, 2015Hi nonlinear, If you wish to match the results of multiple signals as closely as possible, you should trade the signals as a portfolio, meaning taking every signal (buy/sell) even if a signal offsets other signals. You'll need the ability to track ...
Need guidance on managing multiple trade signals
-
FXEZ replied Jul 14, 2015Regarding 0.55 Risk/reward ratio being "not good at all, just .55" since it is Risk over Reward, having a low number is good. If you flip it over for Reward/Risk you get 1.8 profit for every 1.0 of loss - this is essentially like profit target being ...
Is this a good system long term?
-
FXEZ replied Jul 8, 2015Hi PipPirates. Your assumption that the 3 systems are independent is important for this math to work out. This implies that the 3 systems have zero correlation to each other. Let's just work out all the 3 system possibilities: 2 possible outcomes ...
How to find Probability of three Systems
-
FXEZ replied Jun 25, 2015If auto logging off at a fixed hour each day is a problem, see the IBController project which supports staying logged in 24 hours a day. IBController
Interactive Brokers - experiences and opinions
-
FXEZ replied Jun 24, 2015You guys who are considering trailing stops or breakeven stops owe it to yourself to spend some time in this thread: Expectancy Management Spoiler: you may end up rethinking trailing stops / breakeven stops in light of the evidence.
How do incorporate a trailing stop profitably in my trading?
-
FXEZ replied Jun 24, 20151) Make sure your current chart time is the same as the current FF calendar time. 2) What is the "right hour" for a market move? Must the market move at a particular hour? Is news the only driver of price? 3) Is buying / selling activity that ...
Market moves before the news event happens?
-
FXEZ replied Jun 23, 2015Your analysis looks correct for this toy example. How different are your results from a live demo test? Is it due to extra trade costs for executing the 2 pairs instead of the 1 pair? A note of caution regarding applying this analysis using ...
Understanding currency triangulation
-
FXEZ replied Jun 17, 2015This isn't a good definition of random for this context. Try this one instead: In other words, non-skewed or non-trending. When the market trends you tend to either get one or both of the following: more up/down bars and/or larger up/down bars. The ...
Forex Dream Chasers - The 90% rule of randomness
-
FXEZ replied Jun 14, 2015Hi Sis.yphus, Yes I did some studies on sequences a while back and had similar results. From memory, my findings were that: 1) Higher frequency sequences had too small an edge to be worth trading. This is true even if you use a long time frame chart ...
Higher Edge within a Single Candlestick
-
FXEZ replied Jun 3, 2015Thanks for this great summary. It makes a lot of sense. Can this be thought of in terms of the bias/variance tradeoff as well? It seems that the optimization process naturally tends to produce low bias (high performing) systems. Additionally the ...
Machine Learning with algoTraderJo