Statistical analysis of trades

Discussion in 'Trading' started by Richbynature, Dec 15, 2010.

How do you analyse your results?

  1. I don't analyse my results at all

    1 vote(s)
    7.1%
  2. I check to see if I am in profit

    2 vote(s)
    14.3%
  3. I review all my trades to spot patterns

    2 vote(s)
    14.3%
  4. I use formal statistical methods to measure my success

    9 vote(s)
    64.3%
  1. think many people use the 'Account Performance' utility that may come with the
    charting/order entry program they use, for instance NinjaTrader has one, and I have
    a free program - 1 account only - that does similar for Oanda or MetaTrader trades
     
  2. Wallace,

    I have a similar function on the IB account. Would you like to give any details about how you use it?
     
  3. hi Richbynature, I really don't use them much and not familiar with what they can
    analyze, guess a lot of people are only interested in if they're profitable or not

    I'm not math oriented and couldn't use Excel to save my life but you may find this
    site interesting: http://www.financialwebring.org/gummy-stuff/gummy_stuff.htm
     
  4. There is a good book on this topic called Trading Risk by Kenneth Grant. My trading strategies are based on backtests, but I've found real life to be more difficult than backtesting, and so an analysis of my actual trades is essential.
     
  5. Something like

    Position Size (% of capital risked)

    Average Entry Price / Average Exit Price =

    total gain or loss in % of capital risked


    Other interesting things:

    Available Liquidity (so you know if your trading strategy scales)

    correlation to other trades in portfolio

    analysis of runs (wins or losses in sequence)

    equity curve analysis & Average drawdown during a specific sequence of runs

    If you have enough data you could generate allocation strategies on the basis of past trades. Its risk control of portfolio exposure to directional / volatility / liquidity and other variables. Good for regime shifts in which the historical data would help the portfolio even if the actual trading event is not triggered yet.
     
  6. At the beginning of the year it was my focus to set up my spreadsheet to better track my trades. I trade several strategies across a few accounts and now I can analyze data for any strategy or any account, or overall, for any timeframe (I set up input fields for start and end date). So I can see how short strategies do during bull runs, or long strategies during bear runs, or how my trading results this year compare to last year, etc.

    Some of the stats I look at: entry slippage vs. entry signal, exit slippage vs. exit signal (I hold overnight and my signals are based on end of day), commissions, Avg win, avg loss, % wins, Maximum adverse excursion, # Trades, $/Trade, %/Trade, %/trade/day - this last one is very useful for comparing strategies.
     
  7. Redneck

    Redneck

    While learning to trade – it’s probably a good to analyze trades/ setups/ longs/ shorts/ everything/ anything/ whatever – as it could be beneficial / provide insight


    As a trader – I only care about ROCU (return on “net” capital utilized)… Net being less any buying power



    Goal of trading is to make money – ROCU gives insight on how well that money is being managing (ex; $2K a day on $1M…, aint the same as $2K on $25K)


    imo

    RN
     
  8. ugz

    ugz

    statistical measurement is meaningless unless you follow a standardized process, i.e. same entry/exit, capital for all trades like a robot