Bodger Breakout Method

I have a new system but I need to openly ask for help in testing this one. I have employed numerous testing methods, but I don't have the data necessary to do a full-bore backtest with this.

My system is a simple breakout strategy. It works like this:

1) Use a daily timeframe for your chart - candles, bars, whatever you're comfortable with. I will use GBPUSD for this illustration.

2) Set a 1000-period SMA on the chart (yes, I said a THOUSAND). If you don't have a 1000 day's worth of data in your system, then set the MA to the highest value possible that will still give you a value on screen.

3) Look at the chart at the very beginning of a new candle/bar. You will enter ALL of your orders for the day at this time.

4) The 1000 SMA shows you the LONG term trend. If the previous day closed above the SMA, you are looking to buy. If the previous day closed below the SMA, you are looking to sell. In other words, you are looking to trade with the (very) long-term trend. GBPUSD is currently above the 1000 SMA, so we will be placing buy orders.

5) Write down the day's opening price. For illustration, we will use 1.9984 for GBPUSD.

6) Open 5 stop orders like so:
Buy 0.1 GBPUSD @ 2.0039, SL @ 1.9939
Buy 0.1 GBPUSD @ 2.0049, SL @ 1.9969
Buy 0.1 GBPUSD @ 2.0059, SL @ 1.9999
Buy 0.1 GBPUSD @ 2.0069, SL @ 2.0029
Buy 0.1 GBPUSD @ 2.0079, SL @ 2.0059

7) At the very end of the day (just prior to the close of the candle/bar), you close any open positions, regardless of whether they are positive or negative.

Notice that this is a feathering technique. It feathers into the position from 0 to .5 lots in increments of 0.1, and it feathers back out of the position in an inverse manner.

The first stop order is placed 55 pips above the opening price. The other orders are triggered on each successive rise of 10 pips, until you have a total of .5 lots open when the price reaches 2.0079.

If the price then reverses, you will be gradually stopped out of the entire position, with the first lot dying at 2.0059 and each successive lot dying when the price retreats by another 30 pips.

"55" was not picked as an arbitrary number. Almost all daily bars/candles experience some retracement before reversing course and heading in their "true" direction for the day. 80% of all GBPUSD daily candles/bars since 1998 have experienced a daily retracement of 55 pips or less before heading in the day's true direction. This means that once the day's price has travelled 55 pips, there is an 80% chance that this represents the true direction for the day. This method can be applied to any pair, but the value of 55 must be tweaked for each pair. The exact values are as follows:

USDJPY: 43
EURUSD: 45
USDCHF: 62
USDCAD: 41
AUDUSD: 29
EURGBP: 23
EURJPY: 55
GBPJPY: 78
EURCHF: 27
CHFJPY: 28
GBPCHF: 81
EURAUD: 57
EURCAD: 49
AUDCAD: 33
AUDJPY: 31
NZDUSD: 28
AUDNZD: 39
CADJPY: 30

Under this strategy, results will typically fall under four categories:

1) The price moves against you at, or shortly after, the start of the day. This is a great scenario. It means that none of your orders were ever triggered, so your exposure to this pair is nil.

2) The pair moves strongly in your direction, and then strongly reverses. This is the nightmare scenario. It means that the price moved far enough to trigger some or all of your orders, and then reversed against you. Although this is painful, at least you are protected by the series of progressive SLs.

3) The pair moves strongly in your direction and closes strongly in your direction. This is big-win scenario. It means that all of your orders were triggered and will all be in the money by the time that you close the positions at the end of the day.

4) The pair moves strongly enough to trigger some or all of your orders, but then wobbles around for the rest of the day. Maybe the price closes lower and you take a small-to-midsize loss. Maybe it still closes higher and you realize some profit. Depending upon the price action of the day, one or two of your orders can get stopped out but you can still end up positive for the day.

This is where testing gets tricky. You can't just look at daily bars to figure out what would have happened to your positions. You really need to be going through the data minute-by-minute or, preferably, tick-by-tick, to know for certain which orders would have been opened, which ones would have been stopped out, and which ones would have closed positively.

Can someone here please help?

My first day trading this strategy was Friday. I had placed 5 stop orders (in the manner described above) on 19 different pairs. (Yes, it does take a while to enter 5 different stop orders for 19 different pairs, but it is the only real work that needs to be done with this system all day. After that, you can turn your monitor off.)

By the end of the day, I had trades triggered on 13 different pairs. I believe this made it a very heavy trading day for this system. This does not mean that my exposure to all 19 pairs was identical. Based on my feathering process, I was exposed to some pairs (EURCHF) for a mere 0.1 lots. I was exposed to other pairs (CADJPY) for a full 0.5 lots.

Overall, I lost $50 on Friday. I put my results so bluntly in dollar terms, because the traditional view of measuring pips can be very misleading in this system. Technically, you can have as many as 5 different trades compiled into a single position against a given pair. If you count each one of those trades individually, it makes the system look like it is 5x more successful/problematic than it really is.

It's also worth noting that in true recognition of Murphy's Law, my first day trying this system with real money (albeit, miniscule 0.1 lots) turned out to be the classic "nightmare scenario" for this system. The US GDP numbers showed poorly in the morning, which drove a number of pairs high enough to trigger my orders. Then the better-than-expected US consumer sentiment numbers came out later in the morning, driving prices back down. I saw a $40 profit reverse into what at one time was an $80 loss. Of course, this also triggered a lot of my SLs as well.

Another note: If you use a broker with carryover, the daily nature of this system can be turned to your favor for incremental gains. I enter all orders at the beginning of the day's bar/candle (5:00 PM EST with my current data provider). Any positions triggered during the day are to be closed at the very end of that day. I use MB Trading, and they charge carryover at 5:00 PM EST. So if I am sitting on a position that entails a positive carryover, I wait to close it until 5:05 PM EST. All of my positions holding negative carryover get closed just before 5:00 PM EST.

Obviously, this is a very small bonus and comes nowhere close to compensating for a bad trading day. But if you are going to close your positions right at the time that your dealer charges carryover, why wouldn't you selectively hold onto those ones which actually yield a small payment?

If your broker doesn't accept micro (0.1) lots, then you probably need to trade fewer pairs. On any given day, 1/4 - 1/2 of your orders will never be triggered. But on the pairs that DO trigger orders, they could trigger as many as five different orders. This means that you could get overextended very quickly in a big market movement.

If you do have a broker that accepts micro lots (or if you're OK with large exposure), then I do think it's a better idea to trade more pairs. If you only concentrate on one pair, or a few pairs, there will be many days with no activity whatsoever.

If you wanted to provide yourself with maximum diversification without multiplying your exposure to any one currency, you might limit yourself to pairs that don't replicate the same risk. In other words, don't trade USDJPY, EURUSD, GBPUSD, USDCHF, USDCAD, and AUDUSD concurrently, because they can all be construed as a bet for/against the US dollar.

I can see you put alot of work into this method unfortunatly I can't help you testing but just wanted to hope you the best of luck

When does the day start for you?
00 GMT or it does not matter

Thanks

The "day" starts for me at, I believe, GMT -3. I use the data feed from FXDD and my new bar/candle begins at 5:00 PM EST. 5:00 PM EST is 9:00 PM GMT.

However, if you're wondering whether this effects the accuracy of the system, the answer is "no". If your bar begins a few hours off from mine, it means that you may be in a few trades that I'm not, and you may be out of a few trades that I'm in, but in the long run it should even out.

The key point is that over any 24-hour period, pair X tends to retrace by amount Y. If the pair X has moved by Y amount (the actual values are listed in the first post), then you can be 80% certain that the move is genuine.

Do you have any statistics for how many pips it will travel on average once it reaches 55 pips? I did a very quick test for the month of April and although it came out positive it was only very minor. I think things could be improved with a different approach to taking profits rather than leaving the trade open until the day is over.

I have extensively tested idea after idea for different breakout methodologies and they all seem to end up with very minor profits or break even in the end. However my spirit isn't broken yet. I'd love to come up with something that works extremely well.

Great work, thanks for sharing with us

Unfortunately, no.

In this test, did you account for the feathering in/out of the positions? Did you do the test for all pairs? Or just GBPUSD?

Looking back at all of April, I can eyeball the following results for GBPUSD:

April 2 = All 5 trades triggered / +65 pips
April 3 = No trades
April 4 = No trades
April 5 = No trades
April 6 = No trades
April 9 = No trades
April 10 = 5 trades triggered / +130 pips
April 11 = 5 trades triggered / -196 pips
April 12 = 1 trade triggered / -14 pips
April 13 = 5 trades triggered / +5 pips
April 16 = 2 trades triggered / -58 pips
April 17 = 5 trades triggered / +480 pips
April 18 = 1 trade triggered / -35 pips
April 19 = No trades
April 20 = No trades
April 23 = No trades
April 24 = No trades
April 25 = No trades
April 26 = No trades
April 27 = 5 trades triggered / -10 pips

Total for April = +367 pips

I'm open to any suggestions. My first attempt at this strategy was to place progressive TP levels on each trade. The first trade is the one with the widest berth. It is given a 100-pip SL and, consequently, I gave it a 100-pip TP. I gave the second position an 80-pip TP, the third position a 60-pip TP, the fourth position a 40-pip TP, and the fifth position a 20-pip TP. When I ran this through Monte Carlo simulations, I found that the TP actually LOWERED my overall profits. This happened because the downside is already limited by the SLs. But the upside is only limited by the level that can be reached by the pair by the end of the day. Every once in a while, a pair has a big breakout day that pays very handsomely. If you put TPs on the trades, you lock yourself out of these big profits, which sort of goes against the whole breakout idea in the first place.

But I'm not saying all this to insinuate that I have it all figured out - I don't. Maybe TP levels are required. I don't know.

Sorry Mate, I meant March and it was only Cable I tested. I know it's only one month so don't take my comments the wrong way.

I was thinking along the same lines, progressive or staggered TP's only because I hate seeing trades go well into profit only to retrace somewhat. It looks like you have done some real hard work on this though so I trust what your saying. I will try to do some more testing for you as soon as time permits. Working for a living sucks.

Thinking about April's "results", I went back to the beginning of the year to extend my eyeball test. This is what I come up with:

Jan 2 = 5 trades triggered / +105 pips
Jan 3 = No trades
Jan 4 = No trades
Jan 5 = No trades
Jan 8 = 5 trades triggered / -10 pips
Jan 9 = 3 trades triggered / -147 pips
Jan 10 = No trades
Jan 11 = 5 trades triggered / +285 pips
Jan 12 = 5 trades triggered / +305 pips
Jan 15 = 2 trades triggered / -20 pips
Jan 16 = 1 trade triggered / -89 pips
Jan 17 = 5 trades triggered / +55 pips
Jan 18 = 3 trades triggered / -66 pips
Jan 19 = No trades
Jan 22 = No trades
Jan 23 = 5 trades triggered / -96 pips
Jan 24 = No trades
Jan 25 = No trades
Jan 26 = No trades
Jan 29 = No trades
Jan 30 = 3 trades triggered / -138 pips
Jan 31 = No trades

Total for January: +184 pips
Running Total: +184 pips

Feb 1 = 3 trades triggered / -154 pips
Feb 2 = 3 trades triggered / -168 pips
Feb 5 = No trades
Feb 6 = 5 trades triggered / +155 pips
Feb 7 = No trades
Feb 8 = No trades
Feb 9 = No trades
Feb 12 = 3 trades triggered / -202 pips
Feb 13 = No trades
Feb 14 = 5 trades triggered / +425 pips
Feb 15 = No trades
Feb 16 = No trades
Feb 19 = No trades
Feb 20 = 1 trade triggered / -16 pips
Feb 21 = No trades
Feb 22 = 1 trade triggered / -32 pips
Feb 23 = 4 trades triggered / +8 pips
Feb 26 = No trades
Feb 27 = No trades
Feb 28 = No trades

Total for February: +16 pips
Running Total: +200 pips

Mar 1 = No trades
Mar 2 = No trades
Mar 5 = No trades
Mar 6 = 5 trades triggered / +180 pips
Mar 7 = No trades
Mar 8 = No trades
Mar 9 = 1 trade triggered / -18 pips
Mar 12 = 2 trades triggered / -173 pips
Mar 13 = No trades
Mar 14 = 4 trades triggered / +8 pips
Mar 15 = No trades
Mar 16 = 5 trades triggered / -165 pips
Mar 19 = 1 trade triggered / -28 pips
Mar 20 = 5 trades triggered / +485 pips
Mar 21 = 3 trades triggered / 0 pips
Mar 22 = No trades
Mar 23 = No trades
Mar 26 = 5 trades triggered / +25 pips
Mar 27 = No trades
Mar 28 = No trades
Mar 29 = No trades
Mar 30 = 5 trades triggered / -75 pips

Total for March: +239 pips
Running Total: +439 pips

Total for April: +367 pips
Running Total: +806 pips

Now, having posted those "results" for GBPUSD since the beginning of the year, I will be the first to admit that they are questionable. Here's why: I'm only eyeballing the OHLC from each daily bar to try to figure out how many trades would have been triggered and where they would have closed at. The problem is that this does not account for all fluctuations during the day's activities. For example, under my system, it is possible for the price to close 500 points above the open, but I could still get stopped out on a few of my orders if there was a temporary reversal during the middle of the day. This type of data can not be gleaned from looking at daily bars.

That's why I'm asking for helping in testing this.

Sunday night I opened my Bodger Breakout orders for the 2nd day of live trials on this system. After two days I've already come to a significant conclusion.

On Friday I tried this system for the first time and when my layered positions went into effect I was up by ~$100 (~2.7% of total bankroll) by 11 AM Friday morning. Very shortly thereafter, almost all of my open positions turned against me hard and by the end of the day I was down $50.

On Sunday, I opened a new batch of daily orders and when my layered positions went into effect I was up by ~$150 (~4.2 % of total bankroll) by 11 AM this morning. Very shortly thereafter, many of my open positions turned against me and by the time I closed the positions at 3:30 PM this afternoon, I was down $16 (~0.4% of total bankroll).

Now 2 days of experience does nothing to validate or invalidate a system. But I was struck by the fact that my positions turned against me so close to the same time during the day. I also pondered the fact that others in this thread mentioned the possible need to find a better strategy for taking profits.

Under the original system, I was simply closing all positions at the end of the day's bar/candle. This gave the positions plenty of time to run but virtually guaranteed I'd be giving back pips even on the best of days. After all, the odds are pretty strong that the close of the day's bar/candle will not represent the best closing price for my open positions.

I also started thinking about that 11 AM (EST) time frame. 11 AM may not simply be random. After all, I am trading a breakout strategy which depends upon maximum price movement to be successful. Typically, there are 3 events that happen during each 24-hour period which offer the most predictable price action - the opening of the Asian markets, the opening of the European markets, and the opening of the US markets. By 11 AM EST, 2 of those 3 markets have already opened and closed and the 3rd - the US market - has already experienced the initial surge that is typical of the market's opening. I'm not saying that the rest of the day can't possibly offer more price action, but basic market machinations dictate that the bulk of the day's action is over.

So I spent time today eyeballing hourly charts from the last year in many of the major pairs. When I started chopping the charts into 24-hour periods starting and ending each day at 11 AM EST, I was stunned by what I saw. There were a great many days when the day's largest price action was shortly before or during the US morning. During many of these days, the 11 AM hour represented the peak of that day's price action. Very often, the subsequent hours between 11 AM EST and 5 PM EST only represented lateral price action or, in many cases, retracement.

I also went back to my database and started running statistical numbers. Indeed, the historical data shows that a day's extreme price level is typically reached somewhere between 8 AM EST and noon EST.

With all of this in mind, I am going to alter my methodology as such: I will open all of the same stop orders as previously described. However, rather than holding them open blindly until 5 PM of the following day. I will instead close them early at 11 AM, since this seems to have a strong statistical correlation to the day's maximum range.

Certainly, this will lead to some days when I lose additional afternoon pips. However, my research indicates that I will probably save many more pips that would typically come in afternoon retracements. Although this is hardly scientific, I can certainly look back to the last two day's worth of trading and see that closing my positions at 11 AM would have turned a cumulative loss of $66 into a cumulative gain of $250 - a $316 swing (8.7% of total bankroll).

With that in mind, I have placed my usual 5 stop orders for each of 19 different pairs. I am layering in units at a rate of 0.3 per 10 pips of movement.

I will report tomorrow on my results.

Bytebodger...................one of the best things you just did was to take a close look at the 1 hour charts. Granted, in every 24 hour 'trading day' there will be a high and a low, but as you just observed, there is often a pattern to those extremes; and that pattern is related to certain distinct times during each of the three major market sessions. You seem to be on the right track to modify your system.............all the best with your endeavors!

very nice! i like what i see..but i think we need some improvemts on how to set stops...maybe im not sure.

Thank you for sharing!!

I came to similar conclusions myself whilst testing other systems, however I find that the times can vary depending on what pair we are looking at. This does seem to be a better way of taking profits rather than than setting distinct TP points or running a TS. I will try and back test this on Cable over the past 12 months and see what the results show us.

I'm open to any suggestions on SLs. Currently the SLs are set independently for each layer of the position. The first position is given a 100-pip SL, the second an 80-pip SL, the third a 60-pip SL, the fourth a 40-pip SL, and the fifth a 20-pip SL. My reasoning behind this is that I'm trying to alleviate the "nightmare scenario" that can happen if the price moves far enough to trigger all 5 units in the position but then immediately reverses. With these SLs, I will be progressively moved out of the position. However, I'm not trying to claim that this is definitively the best strategy.

I ran a large number of Monte Carlo simulations on different configurations of SLs. My intention was to get a feel for how different SLs work under random price action so I could eliminate the vagaries of any particular currency pair as a variable. With that in mind, I tried setting all of the SLs to a uniform wide (100-pip) level. I also tried setting them all to a uniform tight (I believe I used 40-pips) level. Under my initial testing, the staggered SLs performed best, but the performance of these SLs may indeed be improved under REAL price action. I'm just not sure.

Any specific suggestions on where to set SLs for each of the 5 layered positions?

Thanks for the encouragement, BB! This is definitely a learning experience.

The times certainly vary, but maybe not by as much as you might think. I wrote a program to figure out how many hours it took during the 24-hour trading session for a pair to meet its maximum price level (positive or negative) from the open. The results are as follows:

Pair / Hours to Apex
USDJPY / 16.2
EURJPY / 16.2
GBPJPY / 16.2
CHFJPY / 16.3
AUDNZD / 16.3
AUDUSD / 16.5
EURGBP / 16.6
EURAUD / 16.6
NZDUSD / 16.6
AUDJPY / 16.7
GBPUSD / 16.8
EURCHF / 17.0
EURUSD / 17.1
USDCHF / 17.3
GBPCHF / 17.3
AUDCAD / 17.3
CADJPY / 17.6
USDCAD / 17.9
EURCAD / 18.1
Aggregate / 16.9

Somewhat predictably, many of the JPY pairs peak earliest - but not that much earlier. It's not like I need to close all my JPY trades at 4 AM EST and then close later pairs throughout the morning/afternoon. It's also important to point out that I don't expect to find some "magic time" when I can always close the trades at their maximum profit level. I am just trying to figure out if there is a more intelligent approach than simply holding all positions until the end of the day.

The numbers above are averages - and averages can be problematic. My program also displays the full frequency of values, showing how likely it is that a pair has reached its peak at any particular point in the day, but I didn't want to clutter this thread with some very long data posts (like my thread on pair retracement values), so I will leave it out for now. If you guys are interested in seeing all the values, please let me know and I'll post them in a separate thread.

As averages, it's important to note that if you really want to be confident that you don't miss a large portion of the price movement, you need to bump these numbers up by a bit - which is how I came to the 11 AM standard.