The full title of this book is The Holy Grail Trading System by James Windsor.
The book writes about James’ (Soultrader’s) experience live trading an intraday trend following system that was developed with a team of people. The system performed very well over 14 months bringing a starting account of £10k to £100k before it experienced a ~50% loss and shut down shortly thereafter.
It is an interesting read, and really illustrates a few truths in the world. In James’ case, his team backtested 3 years of data and went live, doing spectacularly well over the following year. As he himself noted, if the market had trended for 2 more years, he would be made. It is the similar case with some hedge funds that started and caught a bull market for 4-5 years, their great track record would then bring about lots of AUM, before getting into trouble when the market turns. By that time the founders and fund managers would have it made, while the investors that chased after performance would suffer.
Yet if James’ team had done backtesting covering lots of market conditions, they would not trade the system live in the first place. Therein lies the irony, if you know too much and think too much, you might just get stuck in inaction. If you didn’t think so much or know so much, and simply just do, you might very well succeed or fail, nobody knows, just like nobody knows if the market will trend or range, and for how long. I think it is the same with entrepreneurship, people get stuck with inaction when they think too much, and there will always be people who just goes ahead and does it, some succeed, some fail, it’s really hard to tell beforehand. That’s life.
Overall, this book is a good read for any retail traders aspiring to build their own automated trading system that mints $$$.
- Research – 3-year backtest of the system with around 600 trades showed an average return of 20% per month, maximum drawdown of 35%, and a single month of loss.
- Start – Started 1 Oct 2005 with £10k
- Success – Balance went to £100k by end-Nov 06 with only 2 down months out of 14.
- Large loss – Had a large loss of 46% in Dec 06.
- Aftermath – After the large drawdown in Dec 06, leverage was scaled down. The system proceeded to have 12 straight days of winners. Jan 07 return was +61%, Feb 07 was -12%. Had leveraged not been scaled down, the system would have made new highs of over £120k in Jan 07.
- End – Trading stopped by end-Feb 07, with balance of £38k, and having taken out over £40k from the income account.
- Epilogue – The system lost in Feb 07, Mar 07, Apr 07, and would have continued to lose and blow up eventually.
- Don’t double-trade the same money
- £10k was taken out and put into an income account, to be traded using the same strategy but money above certain equity targets will be withdrawn for living expenses.
- Even though the main account has less funds, the risk level per pip did not change. In effect, the £10k was being assumed to exist in both accounts and both accounts were traded as such.
- Keep safe money, safe
- 25% of the account was taken out as safe money. However that ‘safe money’ was taken out and put into the income account to be traded using the same strategy which turned it ‘unsafe’.
- Include all market conditions when testing
- With 3 years of backtest and 1 year of live results, it was still not enough because that 4 years only covered a market uptrend.
- Systems need to have a fail-safe
- A system needs to include rules on identifying when the market conditions are causing the system to fail, and what to do when it’s failed.
- Grail was trading intraday trends on just 1 currency pair.
- Trade trends over longer-term and over many different currency pairs.
- Soultrader (UK) – James
- Angelfx (France) – Angel
- Automan (UK) – Tony, programming
- Madisonfx (UK) – Martin, analytics
- Keres (US) – Kev
- Whispy (UK)
- Fluty (Australia)
- Base system
- Take the price of GBP USD at 8am UK time
- Place buy stop at price + 40 pips, sell stop at price – 40 pips
- If trade entry level is hit, delete second trade order
- In both cases, stop loss 80 pips, take profit 240 pips, trailing stop 60 pips
- Close all trades at 6pm UK time
- Changes during the experiment
- At +1 pip in profit, change stop loss to -77 pips
- Changed take profit to 280 pips
- Changed initial stop to 75 pips
- Changed closing time to 6.59pm UK time on longs
- Changed closing time to 3.59pm UK time on shorts
- Changed robot to place trade only when price is within 10 pips of desired entry price, to hide info from brokers
- Money management
- Started with £10k account with £10 per pip
- The per pip level is based on the high-water mark, i.e. roundup(account high-water mark / 1,000), e.g. £12.5k account means roundup(12.5) = £13 per pip.
- From Oct 05 to Dec 05, if account is in drawdown, the per pip amount remains the same.
- From Jan 06 onwards, during a drawdown period, the rules below apply until a time when the high-water mark is broken, at which time the per pip level gets reset to roundup(account high-water mark / 1,000).
- For each losing day, the per pip level increases by £1
- For each winning day, the per pip level decreases by £1