Indicator Reliability Lab
Indicator Reliability Lab allows you to easily backtest the reliability of indicators on the spot without having to learn to write any script languages. You can select any stock / Forex / or any chartable symbols, and run backtests and highlight trades right in the chart! It is integrated into Interactive Stock Charts application for the ease of use!
What's it for?
A very common mistake a trader makes is to try find ONE method, or ONE indicator that works for all stocks. A much better way is to match the RIGHT indicator(s) for a given stock. Indicator Reliability Lab instantly backtests the currently open chart with your selected indicator, parameters and time frame. Instantly you know the key stats such as profit factor, average return, number of winnin / losing trades! There is also a pre-defined list of indicators (RSI, TR/TD, MFI, CCI, more can be added) on EOD daily charts, and you can see at a glance which indicators work better than others for a selected stock!
I am very excited to announce this feature as it will make your trades much more confident and reliable!
Indicator Reliability Lab requires Master Trader Package of Interactive Stock Charts!
Indicator Reliability Lab Part 1
Indicator Reliability Lab Part 2 (with custom screeners)
For example, you want to see how reliabble RSI(14) < 30 rebound strategy. In other words, how likely does RSI(14)<30 really predict the price has bottomed for a certain stock?
Step 1: Open chart view on a stock (select a stock with at least 3 year history is the best). Find IRL button (icon looks like a lab flask)
Step 2: Specify trade entry condition as: RSI, 14, < 30. For Trade action, select long for buy, short for sell.
Step 3: Hold duration means how long you want to hold the position until exit / calculate gain /loss. Instead of complicating things by having you specify exit condition, this gives you a sense how well the entry condition works given the performance of average hold duration.
Step 4: Scan duration Ideally this should be 1000 - 2000 days which covers the past 5 - 10 years history. If you make it too short, there won't be enough test cases or number of winning / losing trades to give you enough confidence that strategy works.
Step 5: Press Simulate button, it calculates vital statistics and populate trades in the table.
Step 6: You can review the trades if you checked Highlight trades in chart This allows you to see simulated trades in the chart by using scroll left / right buttons
IRL is not in the strict sense of backtesting tool, in that you do not specify entry / exit / stoploss and other extranuous details such as commissions and slippage. Rather, it is a simplified version that gives you a sense of how well a particular strategy performs. You specify Entry condition, Holding period, and Scan duration to give you a sense how Reliable the selected indicator strategy works during the past history, and if the probability is very high >85% it will give you confidence it is likely to work in the future as well!
Indicator / Parameter: Select Indicator or Strategy and its parameters to simulate.
For example, RSI(14) would be RSI for Indicator, 14 for parameters.
Operator selection and Compare valueThis is the condition in which trade signal is triggered.
For example, RSI(14) <= 30 would mean trigger a trade action (long or short) when this condition is met.
Trade: Simulate long or short once indicator criteria / signal is triggered
Hold duration: Number of periods to hold position until closing the simulated position.
Scan duration: How far / long back to run backtests
Profit Factor: Total gain / total loss. You want this number to be greater than 2
Probability: Winning probability calculated as p / (1+p) where p is profit factor
Winning trades: Number of winning trades
Losing trades: Number of losing trades
Profit Factor (pf) and Probability Conversion
Profit factor (pf) = Total gain / total loss
Probability = Total gain / (total gain + total loss) = pf / (pf + 1)
If you are bad at math, the following table will give you a quick reference / idea on key values of pf and probability
Want to combine two indicators?
It is well known fact that when you have multiple indicators giving you the same signals, it is extra reliable and strong! It gives you extra confidence to pull the trigger and make money!
However you have to make sure that two indicators are not heavily correlated to each other. The less correlation they have, or independent the better. We call this ORTHOGONALITY in statistical term. Examples of two orthogonal indicators are oscillators such as RSI + TR / TD, or RSI + Trendline patterns. Oscillators such as RSI, CCI, MACD, Stochastics, are heavily correlated. So combining RSI + Stochastics is a poor choice because it rarely increase the reliability of each indicator on their own.
How to calculate combined reliability?
When you know the reliability of two independent / ORTHOGONAL indicators, calculating combined reliability is as follows:
p = 1 - (1-p1)(1-p2)
For example, probability of winning (reliability) of RSI(14) cross over 30 is 70% on stock A
probability of winning (reliability) of TR / TD buy reversal count > 10 is 85% on stock A
Combined probability of TR / TD buy reversal count > 10 AND RSI(14) cross over 30 is 1 - (1-0.7)*(1-0.85) = 0.955 or 95.5% !!!