A repository for reviewing strategy ideas implemented in QunatCT App. Each idea has the following documents:
- A simply structured text file for introducing the idea and its formulation.
- A pine script implementation for TradingView.
- Some graphs and backtest performance metrics.
The outcome of this work is to show that:
- Strategy ideas used in QuantCT App are profitable.
- The probability of having a winning trade and it's expected(average) return, if using QuantCT Screeners.
In addition, and built from those "Yes/No"s, we will have a valubale knowledge base of many trading ideas that PROBABLY won't work, along with some PROBABLY profitable ones!
We want to answer the following question by means of specific statistical tests:
Does this idea have any predictive power (i.e. profitability in future), with a "High Enough Probability" (HEP)?
The HEP here is 0.05 p-value, 0.95 significanse, above "+2 x Sigma" of histogram of mean returns, or simply 95% probability - All of them are same thing.
To do so, each strategy idea is backtested on BTCUSDT starting from 2018-01-01 (a big and available market crash on Binance exchange) to 2020-12-30.
Test details are as follows:
- All tests have $ 100K on initial cash, and the order amount is 1 BTC.
- A backtest is performed on the data mentioned above. Timeframe is based on what the idea suggests (mainly 1d and 4h).
- Applying "Monte Carlo Permutation" (MCP) test with 10,000 random (i.e. non profitable) tests on same data.
- Giving a
Yes, if the mean return of the implemented idea is above "+2 x Sigma" of histogram of mean returns of those 10K random tests, and aNootherwise.
Right now, you can assess each strategy idea in tradingview, as implemented in QuantCT App, and review its win rate, profit factor, etc., and decide to wether use its signals or not for each coin.
In future releases, I will integrate such insightfull informations into the QuantCT Screeners, so that you don't need to leave the app.
The trading/strategy idea is merely Enter/Exit rules, without any risk/money management and without any optimization. If the bare idea is profitable, it has merit of strategy development and optimization efforts.
So, you should NOT blindely follow screener or alert outputs. at least you should consider proper risk and money management for your trade setups.
I will publish several tutorial videos about how to properly use QuantCT Screeners on QuantCT youtube channel.
Of course NOT!
I have several profitable strategies based on
Noideas! It just means that this implementation, on this instrument, with this timeframe, won't be profitable with 95% probability.Maybe it's got a bad luck based on test data, maybe a simple risk management rule elevates it's performance considerably, or any other reason.
However, it's still a wise decision to concentrate on
Yesones first!
This is an under development service of the QuantCT App!
Not,
YET. Maybe a future feature of the QuantCT App.
Mainly statistical inference.
The book "Evidence-based Technical Analysis".
Because I believe you have the right to know what would be the result of using a screener or a signaling/alerting service!
A screener which only combines indicators without giving any insight on the future results, is both useless and dangerous!
Because there are several dynamic-but-insightless crypto screeners right now.
Also, most traders (if not all of them!) ultimately want to trade profitably, rather than struggling with unlimited possible combinations of insightless dynamic screeners.
- Software Engineer.
- Professional crypto algotrader.
- Instructor of crypto algotrading.
- Algorithmic artist.
- Founder of https://quantinvestor.ir
- Founder of https://algoart-gallery.com
- Founder of https://quantct.com