Friday, 13 July 2012

The "Magic" is in the Methodology

I had a comment about periodicities yesterday. Here's the comment and my response:
I must ask. You don't just choose the product; you also change the range bar settings for the product often? How can ther ebe a consistent backtest for this? Frankly, for those of us that have read the blog for some time, it feels like cherry-picking after the fact.
Anon 02:20, for the ES it's always 0.75 points (3 ticks) for renkos with wicks or 5 ticks for range bars - they trade similarly and keep the noise out. Having said that, the periodicity is not the "magic". I'm trading as a discretionary trader, identifying the trend of my chart and then trading the pullbacks. I don't post all my trades - I trade a lot and in different markets. What I'm trying to do is show a chart of a good trade or set of trades so you can see how these these trading pictures setup - there are over a thousand of them now in the blog - so you can learn to identify them in real time. It's helped many people to CP and that's my goal - to help people trade better.
My focus is to identify the trend on my one chart and then trade the pullbacks. Everything else is just managing the trade. And if I keep it simple then I can't muck it up too much.

The pic below shows Flo trading 30 second bars on the DAX. Pretty difficult to do all day as a discretionary trader . I have a hard wired close target and Flo just grinds it out. This is the simplest of algos and the closest I'd want to get to HFT.

Tip: If you use a close target you can easily have a very high win rate. Then all you need to do in the algo is manage the losing exits so you lose less on those few losers than you make on your 75% plus winners. It doesn't have to be complicated.


  1. "Then all you need to do in the algo is manage the losing exits so you lose less on those few losers than you make on your 75% plus winners."

    I assume you mean you lose less in the aggregate on the few losers than on the 75% winners, since the algo shown seems to have a stop loss beyond what the target is for the profit.

  2. Hello Mr EL

    Thanks for everything.

    I must say that on the Algo you posted however,
    the stop is far larger then the profit or you would have been stopped out of that 15:12 trade.



  3. Guys, you should not be stopped out unless the trend has changed. Price going against you is a function of trade location. If you want tighter stops then you have to forgo many good trades and restrict the entries to "better" trade locations. Stops can be one of the major causes of losing. I have a trading plan that doubles up on trades that pullback a lot and are still in the trend. Lots of ways of slicing and dicing. it all depends on how you want to trade.

  4. E.L.,
    Your last comment is golden, and I struggle with the concept early in my trading (after all the cliche is "never add to a losing trade), before I began scaling into trades.

    May I ask, do you pyramid your scale-ins? For example, the first trade is 1 contracts, the second is 2 contracts, the third 4 contracts, as long as the trend is not going against you?

    Thanks for any information. While I have a methodology based much on your writings (and FT71), I struggle still balance disciplined money management with allowing trades to expand and adding on.


  5. Jeff, I clasify a trade as 100%, 50% or 25% and add according to trade location. The exit point/criteria remains the same. Yes, this is a powerful concept.

  6. Thanks for the info, E.L. A powerful concept, but tricky. Managing losses is so important, and not continuing to add while the order flow is going against you. If I get burnt, it's because I've become hyper-focused on adding more entries, instead of checking the context to make sure my reasons are still valid for the trade.

    For me, in addition to trade type, it also depends on whether I'm leaning against an area where I expect quick rejection, such as an LVN (where I don't scale in), or an area where the entry is more vague, around a VPOC, HVN, etc. I prefer to lean against LVNs.

    Good stuff. I know you've done one blog post about this, but I would always enjoy reading more.