Probably the first ever time I haven’t had a Derived Trade
I have found my optimization efforts demoralizing as i dont get to have the forward test results i was expecting to.
Then I came across this article that explains why and even has a name for my case, overfitting.
Here it is, sharing with the community to get the most out of our endeavour
2 Likes
AUDUSD is a correlating pair on AutoChartist’s relevant section. First time ever this happens!! Let’s see how it unfolds. So far I’m not impressed with AutoChartist’s Favourite picks as my own strategy outperforms a widely trusted broker partner!
Correction day
No confirmation trades by AutoChartist. My AutoChartist confirmations account is on a noticeably much worse performance than the Derived Trades account.
Matter of fact Autochartist account is in the loss -0.12% wnile the Derived Trades account is up by 0.21%.
Starting balance 1million, 1% risk per trade, SL@ Bollinger Bands 1 standard deviation on the Daily, 20MA Simple
TREMENDOUS insight and above all practical steps on how to to reduce overfitting of any trading bot in order to increase its predictive performance, by a trader who actually has verified results on Darwinex regulated by FCA!!!
I got StrategyQuant to better my trend scalpers. Extremely fast backtesting and an endless list of very vital feature.
Makes cTrader optimization I wastd my life on, literally a JOKE!
One more thing out of my brain, in order for some mental RAM to free up, in order for a new idea to be generated!!
Maybe I missed out on how you do it in a previous post but how do you adapt the sentiment data from the various feeds you’re getting in StrategyQuant?
For now I aim to use StrategyQuant to build a robust low timeframe trend following automated strategy that agrees with what has been derived.
ie an NZDJPY long on the 1min tf scalping or something
Aggressively switching the strategy to only focusing on deriving trades from TradingView’s forex screener for the following reasons:
- The original deriving strategy is quite time-consuming. Takes about 1 tedious hour to do and might end up making human errors. Consequently, ended up running it once a day instead of multiple times of at least 3 repetitions daily in order to catch timed trend end exits.
- The TradingView trades deriving has been largely automated by freelancers on google sheets. I only need 10 copy-pastes and there I have the derived trades in 5 minutes. Soon to only need 1 copy paste.
- The prestige of TradingView, the severe simplification, and focusing on the very few will help attract more people to participate in this journey and also accelerate the optimization of the strategy.
How the trades are derived:
I calculate the currency strength of each of the 8 major currencies from the Monthly down to the 30m timeframe on TradingView’s forex screener. Why? This is to ensure a Top-Down Analysis and a slower pace of opening and closing trades rather than if I took into account all TFs down to m1
If AUDCAD is a buy or strong buy the AUD takes a +1. If it is a sell or a strong sell then CAD takes a +1. Why? Because the screener assigns a buy value to the pair, if the base currency, in our example AUD, is stronger than the quote currency, in our example CAD.
If it’s Neutral then we assign the directional value based on the higher timeframe. If it’s the Monthly then we assign the directional value from the Weekly timeframe.
We repeat this +1 process across all 28 major pairs. In the end, we have a currency strength list that looks like the following:
MONTHLY
AUD 7
NZD 6
CAD 5
USD 4
CHF 3
GBP 2
EUR 1
JPY 0
Special Case: If there is an equal between currencies then we take into account their pair’s directional value: If CHF and JPY had both a currency strength reading of 2 then we look at the CHFJPY whether is a buy or a sell and assign the +1 point accordingly which is a +1 for CHF if it was a buy else we assign a +1 to the JPY if CHFJPY was a sell.
So we repeat the same process across all 7 timeframes all the way down to m30. Then apparently we end up having 7 lists of currency strength readings.
If for instance, the CAD is always above, meaning stronger than the GBP across ALL 7 timeframe strength lists, then we have a trade that has passed aka derived trade so obviously open a GBPCAD Sell trade.
SL@3*ATR(Daily,20 SMA) 1% equity risk per trade. Why? It is preferable to set the Stop Loss very far away so it never gets hits unless our minimum exit conditions are met. The phenomenon of having my Stop Loss hit and then having the trade reverse back to the intended direction is a frustration to be avoided with a wide Stop Loss. I have started off with 3 ATRs and 20 SMA as an arbitrary starting point and then we correct along the way. Then we readjust the Max Drawdown % to our tolerance in order to increase profits.
Exit when we run again this process and the CAD is no longer above the GBP across ALL timeframe strength lists. (Flip)
1 Like
Post 55 for explanation
Post 55 for explanation