Big Ben Strategy - optimised

tomorton

Legendary member
8,051 1,180
The Big Ben Strategy is a well-known strategy for trading the London break-out on GBP/USD. It relies on setting bracket trades at both extremes of either the day's initial trading range, or the over-night Asian trading session's range - a buy at the high and a sell at the low. The orders are usually OCO, so that one cancels the other - if the buy is triggered first, the sell is cancelled automatically, and vice versa.

I've been back-testing and am now trialling a couple of variations which might prove profitable.

This involves using only the 0700 bar on the GBP/USD half-hour chart. A buy order is set just after 0730 using the high of the 0700 bar and a sell order at the low: the orders are not OCO.

Typically, the half-hour bars during the London session from 8am onwards are twice the length of the 0700 bar and over-night half-hour bars. Price action through the day can often print a range 5 or 6 times the bar's range. This increase in price movement can be demonstrated graphically by bringing up ATR20 along the base of the M30 chart - volatility drastically increases from the start of the London business day and then falls away to a low in the small hours. The oscillations from peaks to troughs are regular and very sharp - its usually easy to judge the time of day or night simply from the ATR value and direction, confirming the potential edge for traders.

Results of back-testing to follow.
 

tomorton

Legendary member
8,051 1,180
Trading tips -

for each order, set a stop-loss at the opposite extreme of the 0700 bar's range: this means as the first trade is stopped out, the opposite trade is triggered

adjust position size to the same percentage of account capital risked: this does mean the entry to stop-loss distance in pips will vary from day to day, but that's a reflection of market volatility: also it allows performance to be logged and assessed continuously: both orders should be of same size

ideally the 0700 bar range should be "typical" for this time of day: if the bar is unusually wide, it might be best to be cautious and leave off trading this strategy today: if the bar is unusually narrow and "inside" the prior bar, use the prior bar's range for setting orders and stops

close the running trade (if still open) at 2pm: cancel the untriggered order (if still pending) at the same time

be aware that its possible to have both trades stopped out in the same session: adjust your trade size accordingly to allow for your own personal risk tolerance

don't set TP's
 

tomorton

Legendary member
8,051 1,180
Back-test summary -

Back-tested using M30 charts, over 7 months, August last year to end of February - 149 trading sessions.

On 40% of days, Trade 1 (either long or short) triggered and was not stopped out, so Trade 2 was not triggered: results at 1400hrs varied from small loss, less than -1r obviously, to +8r. Average result = +2r.

On 30% of days, both Trades 1 and 2 were triggered and stopped out, making a result for the day -2r.

On 15% of days, Trade 1 was stopped out for -1r, but Trade 2 triggered and went on to make a net gain for the day, covering the -1r loss from Trade 1. Average net result = +3r.

On 15% of days, Trade 1 was stopped out for -1r, and Trade 2 triggered and though not stopped out it failed to cover Trade 1's loss. Average net result = -1r.
 

tomorton

Legendary member
8,051 1,180
End of Week 1 of trial.

2 days showed Trades 1 only triggered, for +3r and +2r.
2 days showed both trades triggered and both stopped out, for -2r and -2r
1 day showed both trades triggered and Trade 2 made: net result for day = +0.5r.

Net result for Week 1 = +1.5r.

Follows the back-test statistics pretty well. Happy with that.
 

LouisaM

Junior member
23 3
This is a great trading technique for a day trader, since it allows you to make a profit, taking into account the peculiarities of the transition from one market center to another in the 24-hour currency market
 
  • Like
Reactions: tomorton

tomorton

Legendary member
8,051 1,180
Trading tip -

While NY is only 4hrs behind the UK, close any running trade and cancel any pending order at 1:00pm UK time, not 2.
 

tomorton

Legendary member
8,051 1,180
Not a bad week for the Big Ben trial - Monday and Wednesday were clearly no-trade days as the 0700 M30 bars were so unusually wide. Tuesday saw a long triggered but it closed at 1300 within the 0700 bar range for a minor loss. Thursday was the money day - long triggered but stopped out and short triggered, after which price fell consistently, for a net result at 1300 of +5r. Friday was also a double-trade day, long first, then short, for a net result of +1r.

Net result for the week +4r.


Edit - Actually, this was a +6r net result for the week. I had initially deducted -1r for each of the no-trade days. Onwards and upwards.
 
Last edited:
  • Like
Reactions: trendie and Dowser

tomorton

Legendary member
8,051 1,180
Grrrrr. Short triggered and stopped out, long triggered and stopped out.

But tomorrow's another day....
 

tomorton

Legendary member
8,051 1,180
No trade for me today as the 0700 candle is so long. Though the chart looks pretty bullish so maybe it would have been a winner.

However, the other 4 days of the week have been tradeable for a net gain for the week of +4r.
 

tomorton

Legendary member
8,051 1,180
Brief update in Big Ben trades for March -
18 trading days (2 no-trade days owing to excessive range of 0700 M30 bar)
8 winning days
6 losing days
4 days appr. break-even
total 18 trades at equal risk (r)
best day = +8r
worst day(s) = -2r
net gain = +11r

Of course, if you set position sizes so the the risk per trade, r, is 2% of account capital, this net gain represents a gain of 22% for the month. Happy with that.

Some observations from the last few weeks of excessive volatility - trailing stop-losses constrict gains more than losses, and TP's likewise. I suppose its obvious that in times with extra-wide daily ranges, a strategy that allows winners to run will be more successful.

Can't wait for Monday.
 

trendie

Legendary member
6,452 1,081
Excellent stuff, tomorton. I have been keeping an eye on this idea, just for reference.

One thing that has always fascinated me is outlier/windfall wins: they sometimes come rarely, and if you miss that specific day, it can skew your perceived gains.
Therefore, assume you are going to miss the "big one", and only get averages.
What would your stats look like if you missed the biggest win of the month? the 8R?
What if your 2 biggest days were average? (that is, 11R / 18, times 2)

This is the reason all my win stats currently are switched "off", as I know these are extraordinary windfall times, and mustn't dream the future is always going to be like this.
Hope the above makes sense.

EDIT: I don't want to throw this excellent thread off-course, so I am happy to delete this post, and keep the thread clean, and maybe explore "windfall" on another thread.
 

tomorton

Legendary member
8,051 1,180
Excellent stuff, tomorton. I have been keeping an eye on this idea, just for reference.

One thing that has always fascinated me is outlier/windfall wins: they sometimes come rarely, and if you miss that specific day, it can skew your perceived gains.
Therefore, assume you are going to miss the "big one", and only get averages.
What would your stats look like if you missed the biggest win of the month? the 8R?
What if your 2 biggest days were average? (that is, 11R / 18, times 2)

This is the reason all my win stats currently are switched "off", as I know these are extraordinary windfall times, and mustn't dream the future is always going to be like this.
Hope the above makes sense.

EDIT: I don't want to throw this excellent thread off-course, so I am happy to delete this post, and keep the thread clean, and maybe explore "windfall" on another thread.

Thanks for this interesting reply and following this idea. Its such a well known and long-used strategy I thought it just had to be worth a second look.

Size of winners is very relevant. Across the 8 months I have looked at in detail, the winning days outnumber the losing days by only 75:71. But the net gains amount to +86r. The r:r potential is highly asymmetrical. The worst possible result on a losing day is -2r: but there is theoretically no limit to the size of the winning day's gains. I limit the duration of the running position to 1400hrs (1300hrs while there is only a 4hr time gap with NY). So it is very important to peg the stop-loss and not trail it, and not to use a TP: these will cut gain more than they will cut risk.

It is typical for an extraordinary win of +4r or better to occur to occur at a rate of 2 per month on average. Bearing in mind this represents an achieved r:r on the winning trade of 1:4, that's good going.
 
  • Like
Reactions: trendie

tomorton

Legendary member
8,051 1,180
That's the week ended and also March, so another month's statistics added to the spreadsheet.

20 trading days
8 days with a single order triggered: gain +17r
7 days on which first order triggered and stopped out, triggering second position: gain +2r
5 days on which both orders triggered and both stopped out: gain -10r
Net gain for month = +9r

Not the best result for a month but not bad considering the wildness of the forex market in March.
 
AdBlock Detected

We get it, advertisements are annoying!

But it's thanks to our sponsors that access to Trade2Win remains free for all. By viewing our ads you help us pay our bills, so please support the site and disable your AdBlocker.

I've Disabled AdBlock