The performance of trading strategies based on Internal Bar Strength (IBS) has been declining in recent years. In this article we include backtest results for SPY ETF and a large cap stock portfolio that demonstrate the declining performance. We also discuss the possible reasons of performance degradation.
IBS is a simple indicator defined as follows:
IBS = (Close – Low)/(High -Low)
It is rather amazing how this simple indicator has generated substantial profits when used in mean-reversion mode. I remember an older trader in the office of a CTA I knew in the mid 1990s who was calculating this indicator manually every day after the market close and if there was a signal he was buying S&P 500 futures for client accounts. The following has been a popular strategy using the IBS indicator:
Buy if IBS < 0.20
Sell if IBS > 0.80
Therefore, if the close of the day is closer to the low in comparison to the daily range, the strategy generates an entry signal and exits when the close of the day is closer to the high in relation to the daily range. A signal on September 17 is shown by the green vertical line in the chart below. The entry (green arrow) is at the following open.
Below is a backtest of the simple strategy above in SPY ETF from inception to 09/20/2018. Commission of $0.01 per share is included and equity is fully invested.
CAGR for this simple strategy is 10.08% versus 9.71% for buy and hold. Risk-adjusted performance as measured by MAR (CAGR/Max. Drawdown) is 0.37 for the strategy versus 0.18 for buy and hold due to lower drawdown. There are two main questions here:
- Why has this simple strategy worked well for so long?
- Has the strategy stopped working?
1. In my opinion this strategy has worked well for an extended period of time because the market has been mean-reverting. There is a large class of strategies that use different indicators and all have been highly profitable during the same period of time in SPY ETF. An example is the RSI(2) strategy.
2. Below are monthly returns of the strategy for the above backtest:
It may be seen that from 1993 to 2017 there are only two losing years, -4% for 1994 and -4.7% for 2001. But year-to-date the strategy is down -8.8%. It may also be seen from the equity curve above that there is recently some change in strategy dynamics. It is possible that this simplistic but widely publicized strategy is now in the list of faded strategies by algos and I suspect the good times are over. Also notice that in the last 5 years, there is no gain greater than 10%.
Although substantial gains in the past were made in the 1990s uptrend, the strategy has been unable to match that performance during the recent uptrend. This is possibly due to a crowding effect from the publicity the strategy has received by quantitative blogs in the last 5 years. Also notice how this strategy generated high returns in 2008 and 2011. I believe that was by chance due to special conditions that occurred then.
Portfolio trading example
Below is an example of using the IBS to rank large cap stocks in the universe used in the Premium Stocks reports. There are 84 stocks in the universe from the S&P 100. These are the strategy details:
Market: Stocks from S&P 100 (84 in total)
Time-frame: daily (adjusted data)
Strategy type: IBS, long-only
Backtest period: 01/03/2000 – 09/21/2018
Maximum open positions: 10
Commission per share: $0.01
Position size per stock: Available equity/10
Trade entry: IBS < 0.20
Trade exit: IBS > 0.80
The strategy ranks stocks according to a score equal to 1-IBS and the signals are generated based on trade entry and exit rules for up to 10 stocks with highest score. Position size is available equity/10. Below is the equity curve:
It may be seen that performance was strong during major uptrends but since 2015 there is significant whipsaw in the equity curve. Below are the monthly returns:
Starting in 2015 performance has declined significantly. Although year-to-date performance is positive, we expect further deterioration in the future.
The IBS indicator and associated mean-reversion strategies have worked amazingly well in the past but are now suffering from extensive publicity and changing price action dynamics. We expect further deterioration in the performance of this simple strategy. Note that the parameters used for the entry and exit are nearly optimal and any variations have negative impact on performance on the average. Furthermore, any filter additions also tend to negatively affect performance. IBS-based strategies have worked well in the past because for an extended period of time markets were purely mean-reverting. The current market state is mixed mean-reversion/momentum which has affected this and other simplistic/curve-fitted strategies, such as the RSI(2), in highly negative ways.
If you found this article interesting, I invite you follow this blog via any of the methods below.
If you have any questions or comments, happy to connect on Twitter: @mikeharrisNY
Charting and backtesting program: Amibroker