Observations on profits and losses

Trader's Mindset

Mar 31

I note the behaviors over the years that have tended to make me profits and behaviors that have lost me profits.

These are the behaviors that have tended to make me profits:

  1. Introspecting my internal confidence and carefully weighing it to take my highest confidence ideas. Verbalizing my predictions. Ranking my trade ideas. Honing my intuition and acting on it. Self checking and introspecting my feel for markets, my internal state, etc. Being engaged and energized with the market but disengaged from the results. This is what I describe as “divided mind” outlook.
  2. Synthesizing different types of information and combining it to make novel predictions.
  3. Taking system trades according to the programmed logic. Finding the factors that allow me to take the best opportunities.
  4. Trading markets where persistent bias has been evident, i.e. strong trends or other non technical factors.
  5. Allowing trades time & room to work out and decrease gaming. But also being willing to take a scalp for a small loss.
  6. Doing more research pre-market and running trades and trading according to plan.
  7. Being willing to trade both sides of the market. Reduces risk of serial correlation of losses.

These behaviors have tended to cost me profits:

  1. Over trading. Attempting to too frequently scalp with a tiny stop loss. Retailers, like myself, should probably think of scalping in terms of risk and selectivity but with perhaps a lower frequency.
  2. Getting too big on a single trade or ballooning it until I get stopped out at the low tick.
  3. Adding to a scalp or leaving it on because I plan to make another trade in same direction. Trying to turn a scalp into a home run, as well.
  4. Not practicing discipline. Exceeding daily loss limits. I know that my probability to recover goes down once exceeding certain limits.
  5. Trying to tell myself the market “should be random” when in a trade and read says otherwise.
  6. Trading when I defined regions where the market is “out of bounds”.
  7. Executing too many trades in a narrow window or on the same idea.
  8. Summit fever. Getting too excited about going live after being restricted to paper trading and thus making errors I wouldn’t normally make.
  9. Placing low confidence trades.
  10. Trying to trade at the very limit of my abilities. If I am trading at a level well within my abilities, there is less difference between sim and live. But, if I am trying to trade at a level I can barely maintain then the risk of not being able to replicate in the live is greater.

These are behaviors which have considerable promise but can be very difficult in practice:

  1. Dynamically sizing or using leverage based on confidence. I think for small accounts constant sizing might be best because the minimum risk per trade is often already at or above maximum. On the other hand, sizing up more confident trades can work. Another form of dynamic leverage involves attempting to dynamically vary the leverage while in a trade and this can work too but it can backfire. The risk with the last method, in particular, is when the trade goes against one then it can create larger losses.
  2. Attempting to learn from losses on the same day to make better trades. I used to believe I could do this: however, I am less decided. I suspect that advanced machine learning algorithms are eroding the ability to do this in the simplest way but it is still promising with more sophisticated methods. In any regards, this type of idea really needs > 1 R trades to work or dynamic leverage because if you are taking mostly 1 R trades then at best you will get back to break even. On the other hand, changing leverage makes it more difficult to ascertain if one has an edge.

About the Author

The author is passionate about markets. He has developed top ranked futures strategies. His core focus is (1) applying machine learning and developing systematic strategies, and (2) solving the toughest problems of discretionary trading by applying quantitative tools, machine learning, and performance discipline. You can contact the author at curtis@beyondbacktesting.com.

>