Thursday, April 26, 2012

Back to basics backtest - quality inside bars (OBSOLETE)

4 May 2012: Please note this post is now obsolete.

Results of my recent backtest on inside bars. 

Trading system:
- Place stops at either end of inside bar to trade breakout
- SL = opposite end of inside bar

- inside bar must be in top or bottom 50% of previous candle
- trade breakout in NEXT candle only

Here is a visual example:

Backtest Results

Six currency pairs were tested on the daily timeframe from 2004 to 2011. 257 trades were compiled. Trades with 1:2 risk:reward seem to be the most practical. 1:3 and 1:4 yielded slightly superior returns, but would require you to hold the trade for much longer.

I didn't include the overnight swap rate as it would be near-impossible to calculate for each trade.


- The longest losing streak was six straight losses.
- Inside bars tend to fail breaking out in ranging markets ("stuck in traffic"), flat 21 EMA.
- Seem to perform better at new swing points in "virgin" territory.


  1. Hi

    Interesting approach to filtering inside bars - and more power to your backtesting persistence!

    You might be interested in post 3 on this page
    I am trying to contact Byn to get details on entries, stop loss and profit target used. It may be that 2009 was a lucky one off but it seems worth looking into.

    Might also be interested in the exit strategy used here - scaling out at different fib extensions

    Inside bars, being a universal thing means this shoudl work on lower timeframes giving more trades in a period of time.

    These are also relatively simple strategies and can be formed into metatrader expert advisors for automated trading.

    Can you clarify the results a bit - not sure exactly what you mean by % return on risk per trade - maybe win%, number consecutive looers and other parameters or just post the trade list (entries, exits and inital stop loss) and we can generate our own favorite metrics :-)

  2. Thanks for the link, Strathmore. I'll take a closer look at Byn's results. It's good to swap ideas and backtesting results with other traders.

    Yeah, I'm not sure what metrics other people use to measure "success". I probably should post win% and R:R as well and let people judge on that. Return on risk means precisely that, the return on the amount you risk.

    Example: if your SL on a trade is 20 pips, then you're risking 20 pips. If return on risk is 10%, then you can expect to profit 2 pips per trade (10% of 20 pips). Hope this makes sense. I suppose this value can be called different things like ROI (return on investment) or EV (expected value).

    I've been backtesting inside bars on the 4HR TF as well. I've only backtested three pairs so far from 2010 to early 2012, but the results are looking good so far. Will post more later. Thanks for visiting! :)

  3. Thanks Kevin.

    The system looks promising and I do like the whole concept around trading inside bars

    I would like to try and build a metatrader ea around this - which I would be happy to share with you if it can be done. It would make both backtesting and trading it a little less painful!

    If you are interested could you post or email me
    - a summary of the trading logic you used for the backtest and
    - the list of trades (entry, stop loss, exits, entry and exit dates / times) so I can construct some of the metrics I like to look at myself.

    strathmoreraven at yahoo dot com

  4. No worries Paul, I'm not skilled in creating EAs so will appreciate your help. I'll send a copy of my spreadsheet to you shortly.


Note: Only a member of this blog may post a comment.