Bodger Breakout Method | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

bo7a method... method for GBP/JPY 205 replies

Trend Breakout/Breakdown Method 37 replies

Another London Breakout method 46 replies

Simple Breakout Method 31 replies

Bodger SAR Method 46 replies

Attachments: Bodger Breakout Method
Exit Attachments

Bodger Breakout Method

  • #21
  • May 1, 2007 8:07am May 1, 2007 8:07am
fascinating stuff..................

Quoting Bytebodger
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:

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.
  • #22
  • May 1, 2007 8:27am May 1, 2007 8:27am
Hello,

Sounds like a good system for pyarmding into a trend. I had a quick read of the system, and here is something that just came to mind for Stop losses

in your orginal post you had:

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

I notice that there are no take profits because you close out all trade regardless of the position at the end of the day.

here is a suggestion using your above example...

lot 1) Buy 0.1 GBPUSD @ 2.0039, SL @ 1.9939
lot 2) Buy 0.1 GBPUSD @ 2.0049, SL @ 1.9969
lot 3) Buy 0.1 GBPUSD @ 2.0059, SL @ 1.9999
when you enter the third position, move the stop loss for lot 1 to 2.0039 (Breakeven)
lot 4) Buy 0.1 GBPUSD @ 2.0069, SL @ 2.0029
when you enter the fourth position, move the stop loss for lot 1 to 2.0049 +10pips) move the stop loss for the second position to breakeven
lot 5) Buy 0.1 GBPUSD @ 2.0079, SL @ 2.0059
when you enter the fifth position, move the stop loss for lot 1 to 2.0059 +20pips) move the stop loss for the second position to +10 move the SL for the third position to b/e

in other words, you are using a 20 pip trailing stop loss...

this will help protect some of the profits, if the trades go up, 20/30 pips, but then retrace...

what do you think ?

anthony
  • #23
  • May 1, 2007 10:10am May 1, 2007 10:10am
Quoting anthonyrae
Hello,

Sounds like a good system for pyarmding into a trend. I had a quick read of the system, and here is something that just came to mind for Stop losses

in your orginal post you had:

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

I notice that there are no take profits because you close out all trade regardless of the position at the end of the day.

here is a suggestion using your above example...

lot 1) Buy 0.1 GBPUSD @ 2.0039, SL @ 1.9939
lot 2) Buy 0.1 GBPUSD @ 2.0049, SL @ 1.9969
lot 3) Buy 0.1 GBPUSD @ 2.0059, SL @ 1.9999
when you enter the third position, move the stop loss for lot 1 to 2.0039 (Breakeven)
lot 4) Buy 0.1 GBPUSD @ 2.0069, SL @ 2.0029
when you enter the fourth position, move the stop loss for lot 1 to 2.0049 +10pips) move the stop loss for the second position to breakeven
lot 5) Buy 0.1 GBPUSD @ 2.0079, SL @ 2.0059
when you enter the fifth position, move the stop loss for lot 1 to 2.0059 +20pips) move the stop loss for the second position to +10 move the SL for the third position to b/e

in other words, you are using a 20 pip trailing stop loss...

this will help protect some of the profits, if the trades go up, 20/30 pips, but then retrace...

what do you think ?

anthony
I think it sounds like a logical approach, although I wouldn't personally deploy it. My main issue is merely one of manageability. I'm setting stop orders on 19 different pairs. Once orders start to get triggered, I could have as many as 12 or more pairs with open orders. This means that I don't want to employ a process that requires me to manually shift SLs - it would become impossible to manage. I'm also enamored of the idea of being able to set orders and then leave them alone for a long time without monitoring. Personally, if I wanted to set in TP levels, I would use the upper end of a TTO. That would give me a SL and a TP level.

But this isn't a criticism. Your approach might very well work for some - it just doesn't suit my trading style.

Of course, all of this is moot if I can't get this system to make me some money - see my next post...
  • #24
  • May 1, 2007 10:34am May 1, 2007 10:34am
Well, today was the first time that I employed my tweaked strategy - closing all orders at 11 AM. I don't think there's a problem with closing orders at 11 AM, but it was just a rough day for my system in general.

You see, my process is to look to make trades that breakout in the direction of the long term trend. Long term, the dollar has been very weak against most currencies, so generally speaking, I am always looking to go short on the dollar and long on converse currencies. This means that my strategy is fairly sensitive to USD movements. (Note: You could change this approach for your own trading - my system does not preclude you from tailoring your pairs/positions to ensure that you are not over-exposed to any single currency. It's a personal trading choice.)

This morning things were cooking along just fine. The open of the US market was kicking off a number of my trades and things were going OK until the ISM numbers came out at 10:00 AM EST. These were surprisingly rosy for the US economy and, as such, many of my positions which had recently been kicked off by the usual downward trend in the dollar suddenly reversed. The worst example was GBPUSD, which climbed as high as 2.0073 around 9:30 AM EST and then fell off the edge of a cliff during the next hour.

The only saving grace is that I stuck by my plan to close all positions at 11 AM sharp and before I did, many of the pairs rallied and cut my potential losses considerably.

So these were the results:

Buy 0.3 AUDJPY @ 99.48
Closed @ 99.05

Buy 0.3 AUDNZD @ 1.1164
Closed @ 1.1176

Buy 0.1 AUDUSD @ .8325
Buy 0.2 AUDUSD @ .8326
Closed @ .8273

Buy 0.3 CADJPY @ 107.98
Closed @ 107.93

Buy 0.3 EURCHF @ 1.6500
Buy 0.3 EURCHF @ 1.6510
Closed @ 1.6510

Sell 0.3 EURGBP @ .6803
Closed @ .6802

Buy 0.3 GBPCHF @ 2.4218
Buy 0.3 GBPCHF @ 2.4226
Buy 0.3 GBPCHF @ 2.4235
Closed 0.3 @ 2.4187
Buy 0.3 GBPCHF @ 2.4248
Buy 0.3 GBPCHF @ 2.4254
Closed 1.2 @ 2.4273

Buy 0.3 GBPJPY @ 239.73
Closed @ 239.33

Buy 0.3 GBPUSD @ 2.0050
Buy 0.3 GBPUSD @ 2.0059
Buy 0.3 GBPUSD @ 2.0070
Closed 0.3 @ 2.0010
Closed 0.3 @ 1.9980
Closed 0.3 @ 1.9991

Buy 0.3 NZDUSD @ .7429
Buy 0.3 NZDUSD @ .7439
Closed @ .7404

Sell 0.3 USDCAD @ 1.1054
Closed @ 1.1095

Overall, I lost $101.64 on trades and paid $8.37 in commission, for a net loss to capital of $110.01 (3.1%).
  • #25
  • May 1, 2007 10:41am May 1, 2007 10:41am
On the bright side:

You may notice that maximum exposure to any pair would have been 1.5 lots (0.3 x 5) and I did not in fact have maximum exposure to any of my positions except for GBPCHF - and GBPCHF was my biggest bright spot of the day. I made $24.15 on GBPCHF. That is what the feathering is supposed to do and it performed exactly as planned.

On the dark side:

It's another day of losses. For the last several months I have finished EVERY day (in which I traded) in the red. The winning trades have been few and far between and when they do come they are sandwiched between enough losing trades that the day ends up lower than when I started. I am, as poker players refer to it, "on tilt". Almost all of those losses occured when I was trying other strategies. I have only been using this strategy since last Friday.

On one hand, I think this system will really work. My testing indicates that it should be solid and I believe that the 11 AM exit rule will work very well. Heck, if I had just applied that same rule to my first three days of trading (rather than just starting to implement it today), I would be up by $100. But instead I close another day in a hole. This has been an astonishing losing streak and sometimes it can just be a little difficult to maintain the proper perspective on things. It's like I have the Anti-Midas Touch.

Oh well. Tomorrow's another (trading) day...
  • #26
  • May 1, 2007 11:34am May 1, 2007 11:34am
I find your volatility breakout system concept refreshing.

I know it is frustrating to trade the way you've chosen, but the odds do favor you. The big moves that inevitably occur make the temporary frustration worthwhile.

I know that your concept does have an edge, but it might be less than most traders might be willing to tolerate considering the potential risks.

Since I also use a volatility breakout system (VBS) on forex, I've found out that the more time you can handle in a profitable trade really tends to improve the risk/reward overall.

Yes, you might end up having a loser once in a while that could have been a profit if you didn't hold longer. The way you make money with a VBS is to hold winners as long as possible. Profits are even more pronounced when you are using a trend filter like your moving average.

I'm pretty certain the variance of your system would drop if you only entered once per pair. And the profits would go up tremendously if you increased your hold time from 3 to 6 trading days, though accuracy would drop significantly.

Best wishes on a progressive idea.

TRT
  • #27
  • May 1, 2007 2:08pm May 1, 2007 2:08pm
Quoting TheRealThing
Since I also use a volatility breakout system (VBS) on forex, I've found out that the more time you can handle in a profitable trade really tends to improve the risk/reward overall.

Yes, you might end up having a loser once in a while that could have been a profit if you didn't hold longer. The way you make money with a VBS is to hold winners as long as possible. Profits are even more pronounced when you are using a trend filter like your moving average.

I'm pretty certain the variance of your system would drop if you only entered once per pair. And the profits would go up tremendously if you increased your hold time from 3 to 6 trading days, though accuracy would drop significantly.

Best wishes on a progressive idea.

TRT
Thanks for the encouragement and the advice! Your thoughts will lead me to a new round of testing and research tonight. This definitely gets me to thinking about potential exit strategies...
  • #28
  • May 2, 2007 10:23am May 2, 2007 10:23am
It was a pretty slow day for Bodger Breakouts. I closed my trades at 11 AM and here are the results:

Buy 0.3 CADJPY @ 108.15
Buy 0.3 CADJPY @ 108.26
Buy 0.3 CADJPY @ 108.36
Closed @ 108.42

Buy 0.3 CHFJPY @ 98.96
Buy 0.3 CHFJPY @ 99.06
Closed @ 98.97

Sell 0.3 USDCAD @ 1.1066
Closed @ 1.1080

Buy 0.3 USDJPY @ 120.29
Closed @ 120.11

Overall, I made $1.50 on trades and paid $2.40 in commission, for a net loss of $0.90 (-0.02% of total capital).

Basically, the day was a push.
  • #29
  • Edited 10:22am May 3, 2007 10:06am | Edited 10:22am
Despite my plea, I haven't heard from anyone who can help me backtest this. My primary concern is that this should really be fully tested with tick data or, at a minimum, minute data, but I have access to no tick data and my minute data only goes back to February, which makes any backtesting almost useless. Nevertheless, I decided that I needed to have some concrete data, so yesterday I loaded up the 5-minute data for all 19 pairs and backtested them against this system. The 5-minute data isn't as ideal as having tick- or minute-data, but it should still be quite accurate and in this manner I can get data back to last December. I came up with some very interesting results.

First, the system, as I have previously defined it in this thread, would have been profitable if deployed since December 12th of last year. The level of profitability is of course, largely dependent upon money management, but it does bring in pips. The pip results were as follows:

USDJPY = +1106
EURUSD = -216
GBPUSD = +856
USDCHF = -526
USDCAD = +538
AUDUSD = +317
EURGBP = +106
EURJPY = +2170
GBPJPY = +3454
EURCHF = +229
CHFJPY = -569
GBPCHF = +1640
EURAUD = -298
EURCAD = -761
AUDCAD = -386
AUDJPY = +2547
NZDUSD = +117
AUDNZD = -394
CADJPY = +1756
-----------------
Total = +11686

Before we go any further, I want to make it very clear that the +11686 number needs to be taken with a grain of salt. I wrote my backtesting program to simply tally all pips. With my feathering process in place for opening positions, this has a tendency to exaggerate both wins and losses. Take the following hypothetical Bodger Breakout trades as an example:

Buy 1 USDJPY @ 119.50
Buy 1 USDJPY @ 119.60
Buy 1 USDJPY @ 119.70
Buy 1 USDJPY @ 119.80
Buy 1 USDJPY @ 119.90
Sell 5 USDJPY @ 120.00

The "raw" pip tally on these trades is +150 (+50 on the first lot, +40 on the second lot, +30 on the third lot, +20 on the fourth lot, and +10 on the fifth). However, this price movement, under more conventional trading tactics, would be seen as a 50 pip gain (bought @ 119.50 and sold @ 120.00).

This consideration does not invalidate the results. I think it's clear that this can be a profitable system. However, the results are not as rosy as you might expect if I were otherwise to tell you that I had made more than 11000 pips since December.
  • #30
  • May 3, 2007 10:22am May 3, 2007 10:22am
Quoting Bytebodger
USDJPY = +1106
EURUSD = -216
GBPUSD = +856
USDCHF = -526
USDCAD = +538
AUDUSD = +317
EURGBP = +106
EURJPY = +2170
GBPJPY = +3454
EURCHF = +229
CHFJPY = -569
GBPCHF = +1640
EURAUD = -298
EURCAD = -761
AUDCAD = -386
AUDJPY = +2547
NZDUSD = +117
AUDNZD = -394
CADJPY = +1756
-----------------
Total = +11686
The first thing that smacked me in the face when I saw these results was the relative performance of all yen-related pairs. GBPUSD didn't perform too badly, but GBPCHF was the only non-yen-related pair that put up numbers on par with the yen-related-pairs. Conversely, CHFJPY is the only yen-related pair to show poorly. If we assume that this is a system to be primarily utilized on the yen and GBPCHF, then we get the following performance since December:

USDJPY = +1106
EURJPY = +2170
GBPJPY = +3454
CHFJPY = -569
GBPCHF = +1640
AUDJPY = +2547
CADJPY = +1756
----------------
Total = +12104

That's basically a slightly better performance returned from a lot fewer trades which equals a lot more money. But I'm uncomfortable with the idea that I should only concentrate on the yen and GBPCHF. I mean, what is it about these pairs that would somehow make them perform better under this system?

I have always been very skeptical anytime someone says, "I have a great system, but it only works on Pair X or Pair Y." If you backtest something and it tanks against 18 pairs but then performs spectacularly against a single pair, then the odds are very high that the positive return on that single pair was a statistical anomaly - basically, luck. Having said this, I think it's more likely that certain systems work better for ALL pairs related to a given currency, rather than being confined to a specific pair. If a pair works very well for all pairs related to a given currency, I suppose it's possible that the specific currency encompasses certain characteristics in its price movement that make it a better target for a given system.

Of course, all of this is clouded by the fact that I only have 5-minute data to backtest against back to December 12, 2006.
  • #31
  • May 3, 2007 10:33am May 3, 2007 10:33am
So I'm thinking further about why the yen may work so well on this system. Is it volatility? After all, this is a "breakout" system and the yen pairs typically have a high daily range. Here are daily range values for the pairs:

GBPJPY = 208
GBPCHF = 204
USDCHF = 154
EURAUD = 142
EURJPY = 140
GBPUSD = 137
EURCAD = 128
USDJPY = 119
EURUSD = 109
USDCAD = 103
AUDNZD = 100
CADJPY = 90
AUDCAD = 83
AUDJPY = 80
NZDUSD = 76
AUDUSD = 74
CHFJPY = 71
EURCHF = 66
EURGBP = 55

Well, as you can see, daily range alone doesn't shine much light on this. Yes, some of the yen pairs have high daily ranges, but other yen pairs have relatively low daily ranges and still perform very well. Conversely, a pair like USDCHF has a very high daily range, but performs poorly in the backtest.

So maybe it's the fact that the yen-based pairs have all been trending very steadily upward (meaning that the yen has been falling) since December? After all, my system uses a 1000-day SMA to determine in which direction we are trading. Unfortunately, this doesn't do much to answer the question. The EURUSD has been trending very strongly upward since December, but does not perform well at all in this backtest.

To make a long story short, I'm honestly not sure why all the yen pairs (and GBPCHF) would have performed so well in this system since December, which actually makes me a little nervous. However, the empirical evidence is so striking that I am going to move forward with the assumption that this should only be traded against the yen pairs + GBPCHF. (You could throw in GBPUSD as well. GBPUSD had positive results, just not nearly as positive as the yen pairs.)

Granted, in 6 months, I may be lamenting the fact that the yen pairs performed horribly under this system while the other pairs would have made tons of pips. But with the data available to me, I can't justify continuing to trade all 19 pairs.
  • #32
  • May 3, 2007 11:08am May 3, 2007 11:08am
Once I had all the 5-minute data in my database, I was able to evaluate the empirical difference between SL levels. Remember, under my original guidelines, SLs would be set as follows on a series of hypothetical Bodger Breakout trades:

Buy 1 USDJPY @ 119.50, SL @ 118.50
Buy 1 USDJPY @ 119.60, SL @ 118.80
Buy 1 USDJPY @ 119.70, SL @ 119.10
Buy 1 USDJPY @ 119.80, SL @ 119.40
Buy 1 USDJPY @ 119.90, SL @ 119.70

Put into layman's terms, 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. The idea is that once you're loaded up on all 5 positions, if the price turns hard against you, the positions are sequentially stopped out after every negative turn of 20 pips.

The backtest results displayed in my previous posts did NOT use this strategy. Those backtest results used the threshold as the SL level. The threshold is the number of pips above the day's opening when you start to open your orders. For example, USDJPY has a threshold of 43. This means that if the day opens at 119.07, the first buy order would be placed at 119.50 (Open + Threshold) and subsequent buy orders would be feathered in every 10 pips thereafter.

The original backtest numbers used a feathering strategy like this:

Buy 1 USDJPY @ 119.50, SL @ 119.03
Buy 1 USDJPY @ 119.60, SL @ 119.13
Buy 1 USDJPY @ 119.70, SL @ 119.23
Buy 1 USDJPY @ 119.80, SL @ 119.33
Buy 1 USDJPY @ 119.90, SL @ 119.43

Put into layman's terms, every open position for USDJPY was given a 43-pip SL. The results of this test were listed in the previous post. For reference, the cumulative pip total for this backtest on the targeted pairs was +12104.

What are the results when I employ the original SL strategy (100-, 80-, 60-, 40-, and 20-pip stops)? They are as follows:

USDJPY = +974
EURJPY = +1473
GBPJPY = +2915
CHFJPY = -416
GBPCHF = +392
AUDJPY = +2606
CADJPY = +1371
----------------
Total = +9315

As you can see, employing my original SL technique would have cost me almost 3000 pips since last December. So it appears that the progressive SLs are not optimal.

My next test used a SL value of 1.5 x the Threshold. In other words, since the USDJPY threshold is 43, I gave each position a SL of 65 pips. The results of this test were as follows:

USDJPY = +1256
EURJPY = +2068
GBPJPY = +3662
CHFJPY = -635
GBPCHF = +1640
AUDJPY = +2547
CADJPY = +1756
----------------
Total = +12777

So the 1.5x value did yield more pips, although not a jaw-dropping amount. But what if we don't make the SL a function of the Threshold at all? What if we just set a uniform SL level? Next I set all positions to have a 100-pip SL. These were the results:

USDJPY = +1592
EURJPY = +2082
GBPJPY = +3824
CHFJPY = -313
GBPCHF = +1559
AUDJPY = +3369
CADJPY = +2590
----------------
Total = +14731

This is a more significant increase. It tells me that each of the positions needs to be given ample room to breathe - even if the price action turns against you. Of course, you need to have the stones to take on larger occasional losses, but the net gain is sizeable.

Finally, I tried setting all positions with a 150-pip SL. These were the results:

USDJPY = +1592
EURJPY = +2323
GBPJPY = +4475
CHFJPY = -336
GBPCHF = +1437
AUDJPY = +3330
CADJPY = +2590
----------------
Total = +15411

Once again, net pips increased, but not enough to really justify the added downside risk inherent in a series of 150-pip stops. So I feel pretty confident saying that each position in the feathered stack should have a 100-pip SL.
  • #33
  • May 3, 2007 11:57am May 3, 2007 11:57am
If you've read my flurry of posts, you know that I've now honed in on a smaller subset of pairs (basically, all the yen-related pairs and GBPCHF). I've also come up with more effective SLs (every position has a 100-pip SL). The next thing for me to look at was the threshold.

The threshold is the point at which you start opening positions and every pair has its own threshold based on its daily volatility. For example, the threshold for USDJPY is 43. Therefore, if USDJPY opens at 119.07, the first buy order would be placed at 119.50, the second at 119.60, the third at 119.70, etc. So it's logical to ask whether this is the right level at which we should start opening positions.

The first thing that I did was to set the threshold to 10 for all pairs. This means that if a pair opens at 119.00, the first buy order would be placed at 119.10, the second at 119.20, etc. These were the results:

USDJPY = +2974
EURJPY = +4212
GBPJPY = +11416
CHFJPY = -158
GBPCHF = +4955
AUDJPY = +4574
CADJPY = +3251
----------------
Total = +31224

Whoa...

This result was a complete surprise to me. This was more than 2x as good as any of my prior tests. By giving each position a 100-pip SL and starting to open them as soon as the pair had moved as little as 10 pips, I added ~17000 pips to the bottom line. This looks tremendously promising, but it also brings up a lot of questions/worries.

First, if these results are valid, then it means that this is really just a simple momentum-following strategy more than a true breakout system. Afterall, a 10-pip increase from the day's opening price is rarely considered to be a "breakout". Not that I really care too much about definitions - pips are pips - but I prefer to know the underlying theory behind my trading strategy.

Second, this scenario widens the divide between the haves (the yen-related pairs and GBPCHF) and the have-nots (all the remaining pairs). Let me show you what I mean. If I run this strategy against ALL pairs, these are the results:

USDJPY = +2974
EURUSD = -934
GBPUSD = -1021
USDCHF = -2169
USDCAD = -151
AUDUSD = +1146
EURGBP = +51
EURJPY = +4212
GBPJPY = +11416
EURCHF = +990
CHFJPY = -158
GBPCHF = +4955
EURAUD = -476
EURCAD = -2575
AUDCAD = +36
AUDJPY = +4574
NZDUSD = +417
AUDNZD = -1764
CADJPY = +3251
----------------
Total = +24774

Notice what happened here. When I removed the threshold and started opening positions very shortly after a pair started moving, my chosen pairs (USDJPY, EURJPY, GBPJPY, CHFJPY, GBPCHF, AUDJPY, CADJPY) all displayed a dramatic increase in performance. The pairs that I've already singled out (all the others) showed a dramatic decrease in performance. But I find it hard to believe that such a simple momentum play could work so consistently on the yen pairs and so poorly on almost all the others.

I'm not sure if this is another limitation of my small data set, or if these are truly "legitimate" results.

Very perplexed....
  • #34
  • May 4, 2007 6:04am May 4, 2007 6:04am
I'm finding it a little hard to keep up with you now. During your last few posts you have shown some impressive results on several pairs. This, although always seems great to newbies, scares me, as the back testing almost always seems to be mistaken. Forgive my skepticism though, I can't help it.

Are you still closing all your orders at 1600GMT?

May I ask what program you are using to perform these back tests and where the data came from as well?

I am currently back testing Cable from Apr 06 to Apr 07 and will post results shortly. I am using your threshold as the initial buy and feathering stops 100, 80, 60 etc. Unfortunately I am a busy man and can't devote all my time to this project.
  • #35
  • May 4, 2007 8:12am May 4, 2007 8:12am
Quoting Squash
I'm finding it a little hard to keep up with you now.
Hehehe - sorry to have hit you with a flurry of posts. I just did a lot of backtesting a few nights ago and it was too much data and too many concepts to reasonably fit into a single post.

Quoting Squash
During your last few posts you have shown some impressive results on several pairs. This, although always seems great to newbies, scares me, as the back testing almost always seems to be mistaken. Forgive my skepticism though, I can't help it.
I couldn't agree with you more! The only thing more suspicious than a negative backtest result is a positive backtest result. As I have stated in some of my previous posts, the varying nature of these results gives me pause and I am not presenting this evidence to "make a case" for my system. I am only presenting it as the sum of all data that I currently have available to me. It definitely needs more testing.

Quoting Squash
Are you still closing all your orders at 1600GMT?
Yes. I ran most of these tests with the filter set so that I closed all orders at 11 AM EST (I believe that is 1600 GMT). After I started focusing on the yen pairs, I tried the test with closing all orders at different times but could come up with no better hour.

Quoting Squash
May I ask what program you are using to perform these back tests and where the data came from as well?
These tests were performed using the 5-minute data from FXDD. I downloaded all the data into my own MySQL database. I then wrote a PHP program that went through the data 5 minutes at a time and executed my orders according to the criteria. The PHP program compiles the results for me in my browser.

Quoting Squash
I am currently back testing Cable from Apr 06 to Apr 07 and will post results shortly. I am using your threshold as the initial buy and feathering stops 100, 80, 60 etc. Unfortunately I am a busy man and can't devote all my time to this project.
I understand. Any results you can provide would be very helpful. As you can see from my results, Cable was profitable - although not as profitable as GBPCHF and the yen pairs. I am very curious to see how Cable performed over a longer time period.
  • #36
  • May 4, 2007 9:29am May 4, 2007 9:29am
Quoting Bytebodger
These tests were performed using the 5-minute data from FXDD. I downloaded all the data into my own MySQL database. I then wrote a PHP program that went through the data 5 minutes at a time and executed my orders according to the criteria. The PHP program compiles the results for me in my browser.
Very cool. Thats a lot bloody easier than how I am doing it.

I have results for Cable as stated in my post above, from Apr 06 to Apr 07. This is data from Alpari servers accurate(as can be) to the minute.

Total profit is 3510 pips which is including the spread of course.
The Max draw down at any point was 1038 pips.
Total Orders put through was 390. So we have effectively given 1560 pips to the broker.

Not a bad result, the only discouraging thing would be the rather large draw down. But an average of 293 pips per month is great for one pair, especially considering its a set and forget trading system. This is the great thing about breakout systems as far as I'm concerned.

Now I think I will try and do a years testing on GBPJPY after seeing your results. I think I will try first opening positions 10 pips past the open price and a 100SL set across the board. They looked liked your most promising results. Will report soon.
  • #37
  • May 4, 2007 10:32am May 4, 2007 10:32am
Quoting Squash
Very cool. Thats a lot bloody easier than how I am doing it.

I have results for Cable as stated in my post above, from Apr 06 to Apr 07. This is data from Alpari servers accurate(as can be) to the minute.
Holy freaking cow. Your post led me to do a quick Google search on "Alpari" and I quickly found a page that purports to have the minute data for a huge list of pairs going back to 2004. If that is so, then I can get it all downloaded into my database and automate ALL of this backtesting (as well as backtesting for any number of other strategies). I hadn't found this before.

I would be MUCH more comfortable doing backtesting against all pairs a couple of years. THANK YOU for this clue!!!
  • #38
  • May 4, 2007 7:38pm May 4, 2007 7:38pm
Ah geez. Sorry mate, I thought everyone new about Alpari. Here's a link to the web site. http://www.alpari-idc.com/en/dc/databank.html

Good luck with your testing.
  • #39
  • May 4, 2007 8:27pm May 4, 2007 8:27pm
Quoting Squash
Ah geez. Sorry mate, I thought everyone new about Alpari. Here's a link to the web site. http://www.alpari-idc.com/en/dc/databank.html

Good luck with your testing.
No worries. I just didn't realize that this type of historical data was available. You see, I'm a programmer so I can write my own backtests - IF I have sufficient data to test against. I have daily data going back to the 70s in the case of some pairs, but for some systems, daily data is not sufficient for backtesting. Specifically, the SLs and the feathered openings on this system make the daily data almost useless.

I am currently in the process of importing ~15 million records into my database. Once that has been completed, I will start to post data on this system going back to 2004.
  • #40
  • May 4, 2007 10:55pm May 4, 2007 10:55pm
Bytebodger write

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.

why don't you just open buy and sell (two direction)?
buy is above 55 pips above from open and sell is below 55 pips from open...

0 traders viewing now