Robot Results March 26th to March 31st

Another week gone. Let's see how our 5 demo test robots did this week.

Inside Bar

Last week it made $316. It's made profit (I think) every week since switching it to the 4 hour time frame. Except for this week. Only 2 inside bars formed on EURUSD and 1 on AUDUSD. Some small amounts of profit but two losses on EURUSD of $110 each which left us down $133 for the week. Balance is $4825.

It's quite frustrating to think you have a good system, and then have to wait an entire week to find you've lost money. That's trading for you. 

Pending Reversal Candle

Several weeks of losing- last week $515 down. This one fails back-tests miserably. So I was very surprised when this was one of the few robots to make money this week. Not just make money, but all of it's trades were in profit. It's up $166, with one order currently open for $24.73.

It has a long way to go to recoup it's losses, and I doubt it will get there, but this has been an extremely interesting result. 

Weekly Breakout 2

This used to be the 4 hour breakout box system, which showed some promise. I traded it manually but missed a few entries and lost badly. So I automated it, using my own code instead of the EA that comes with the system.

I still think it's promising. This is not what you'd think looking at this weeks results though, huh? $465 down. 3 big losses as GBPJPY whipsawed and lost $286 and $287. Then another $158 loss on AUDUSD. A few smaller wins and one order currently open, GBPJPY for $254. 

I'm sure the risk/reward ratio is not what it should be. I'm way too optimistic about this system and haven't set things up right, I suspect.

The robot didn't place some trades which I thought it should have. The problem was if a BUYSTOP order is stopped out, but the SELLSTOP is still in place, the code will not place another BUYSTOP even if the price is back in the breakout box region. I shortened the expiry of the orders from a day to 4 hours so hopefully this won't happen. 

Also, I happened to manually move the stop-loss on a couple of orders to near break-even, which ended up being a mistake. Price retraced, closed, then continued on it's merry way, leaving me short many pips. Sigh. It's a problem with longer term systems like this. The tendency to meddle instead of leaving price to do it's thing. 

Daily Close Breakout

Two weeks ago this system made $812. Last week it lost $1140. This week it's down another $497. I decided to switch this to the 4 hour time frame, but then forgot to restart it. Despite the sometimes impressive results, I doubt this system is going to do very well. Balance is $4056.

 Open Close Bot H4

Lastly my old trend following robot, which I keep solely to gauge trends. It had some good weeks a while back, but the last few weeks have not been kind to it. Too much choppiness - this system needs several 4 hour bars to go in the same direction, and this did not happen much this week. 

Last week it lost $587. This week it gives back another $563 with mostly losses and only two small wins. Balance is $4257.

So, another very disappointing week in robots, or rather, an ordinary week. The weeks where the robots are in profit are strange and unusual weeks!

Lifesdream Free Robots

This website has several free robots you can download. There are two categories, Grid and Trend.

Let's look at a couple. The first one is RSI Grid, and according to the back-test results, it's made over $60,000 from a $10,000 account in a year. 61% win rate, averaging $205 per win.

Quite impressive, until I download it and run it myself.

(Back test from 1-1-2011 to 1-1-2012, GBPUSD, 5 minute chart)

I used the rsi_grid_set2 set file from their website, which supposedly yields the $60,000.

Not such great results. My history data may be messed up, and I'm using different dates for the back-test, but the result is wildly different - a $4639 loss.

It's still achieving a nearly 61% win rate, but the average win is only $113.

Also, it only places trades up to the 18th Janurary 2011, when it places a losing trade with a 1.90 lot size and suffers a single $3131 loss. Ouch.


Let's try Momentum Grid, another robot from their grid category.  It's back-test results show a $50,769 profit from a $50,000 account over 13 months. 66% win rate, averaging $100 per win to $103 per loss.

I noticed some of the results links take you to the RSI Grid results, so be careful what you click on.

(Back test from 1-1-2011 to 1-1-2012, GBPUSD, 5 minute chart)

I used the momentum_grid_set1 set file. Again, it does OK until it starts making trades with huge lot sizes a mere 31 days in, and loses most of it's gains. The robot must detect this, and stops trading.

Up until that point, this robot presents a beautiful equity curve. It trades frequently, has a 67% win rate according to my results, and averages 3 consecutive wins to 1 consecutive loss.

Perhaps adjusting the settings or playing with the code to prevent the huge lot sizes would make this a wonderful robot.

I'm impressed that it makes profit on the 5 minute chart, particularly on GBPUSD. I don't think any of my robots are successful with that time-frame and currency pair.


Let's try the Heiken Ashi trend robot, which according to it's first back-test made $19,442 in about 13 months. 62% win rate, averaging $633 per win.

(Back test from 1-1-2011 to 1-1-2012, EURUSD, 60 minute chart)

Yikes. Everything was going well until 19th July, and then it was all down-hill from there. This robot has 5 back-tests posted on the website, and from the looks of the equity curves, it was a struggle to get this thing making money.

My results using the first set file show only a 44.9% win rate, averaging $250 per win, $265 per loss, losing $6591 over the year.


Last one, let's see how Fisher M11, another trend system, does. According to back-test 3, it makes $19441 in a little more than 4 months. 58% win rate, average win of $882.

(Back test from 1-1-2011 to 1-1-2012, EURUSD, 60 minute chart)

Using the fisher_m11_set3 set file, my results are very different from what's on the website. Although the equity curve is very pleasant, on the 14th April it places a trade using a lot size of 7.20 and within 3 hours has lost $6660. The robot then stops.

57% win rate, $1861 loss overall from a $10,000 account.

What's with these gigantic lot sizes? Are they the secret to making profit with these robots? If so, it's not working out very well, at least according to my four back-tests, which all end the same way.

Looking through the robot settings, I can see an option called "Progression", which I think is related to the lot sizes. Turning it off and running this back-test again yields these results:

(Back test from 1-1-2011 to 1-1-2012, EURUSD, 60 minute chart)

33% win rate, $9922 loss. Lasted almost the entire year before running out of capital, though.

Nice website. Nice looking code in the robots, but none of them in the configurations I've tried in my back-tests have yielded positive results.

EURUSD - 2012-03-30

Well, here's my best bet for what's happening with EURUSD this week. A nice orange trend-line which shows a bullish trend starting 3 weeks ago.

(60 min EURUSD chart)

EURUSD has moved above the daily 100 period simple moving average (not shown), and is still above all of the daily Ichimoku Kinko Hyo indicators (not shown). 

Price hasn't reached the February high of near 1.35. I have no real idea what's happening. I can kind of draw a bearish trend-line on the weekly chart (not shown), but it's only approximate, not actually touching several points properly. 

It dipped below last weeks high a few 4 hour bars ago, and currently looks to have tested and failed to reach this weeks high. 

(240 minute EURUSD chart)

There's a lot of long tails on these 4 hour candles this week, which indicates market uncertainty. I don't like the looks of it. 

EURUSD, 2012-03-26

Last week, several technical indicators seemed to show that the EURUSD had nowhere to go but down. So of course, it went up. 

(Weekly EURUSD chart, with last weeks price action in yellow)

Looking at the weekly chart, it all seems so clear. The EURUSD went up. No doubt about it. But during that week, things were not so clear. 

(4 hr EURUSD chart, with last weeks price action in yellow)

Price went up, but not before being indecisive several times. Also, we can see the 100 and 200 simple moving averages cross right in the middle of this price action.

Near the end of the week, a lovely head and shoulders pattern appeared on the hourly chart. 

Longer term, however, price does appear to still be moving downwards. 

 (Daily EURUSD chart with 100 and 200 period simple moving averages plotted)

Although price has again closed above the 100 period simple moving average on the daily chart, it looks like it may run into my hastily drawn bearish trend-line. If that happens, then it's bear city and 1.30 re-test time.

It sure seems like I say that a lot though, which means EURUSD is really still in a consolidation zone between 1.35 and 1.30.

My breakout trading strategy hasn't been working out well, with last week seeing a $60 loss and a $30 win for a total loss of $30. It obviously needs some reviewing and careful attention.

For this week, I will keep a careful eye on this daily chart and the orange trend line, and as always, will be watching other pairs for clues as to where price may be going. It sort of seems like technically, EURUSD is set up for a move south, but the other pairs want to go up, which will probably drag EURUSD higher.

I'll also be keeping an eye out for more consolidation patterns so I can place some trades. I am fast losing hope and heart, but I guess this is one of those things you need to get through to be a Forex trader.

Robot Results March 18-24

A pretty poor week for the 5 robots we've been reviewing for the last several weeks. Let's see how they did.


Inside Bar

Last week $40 profit. This week $316 and a lot more active in trading. It's interesting that this robot's trading activity has picked up, considering no changes to the code and the dismal results of all the other robots reviewed this week. I'm guessing this means a lot of market uncertainty, which this robot was able to capitalize on. Moving it to the 4hr time frame definitely seems to have paid off  (so far). Balance is $4958.


Pending Reversal Candle

Last week I added AUDUSD for this robot to trade, and it did poorly, ending the week with a $180 loss. This week was much worse - a $515 loss, with all of the losses coming from AUDUSD. Current balance is $4906.

This robot was starting to look promising running on just EURUSD.

It's tempting to take AUDUSD off, but even if this robot continues to fail on AUDUSD, we can use the results for comparison with the others robots trading AUDUSD.

Testing takes a long, long time. If this robot does fail, I'll restart the test with just EURUSD again and run it for another few months.


4 Hour Breakout Box

Another horrible week for this otherwise promising system, which I was trading manually. It's frustrating, because even though there were some terrible losses, there were plenty of opportunities for good trades which should have at least made it break even.

The fault is all mine, of course, for not keeping a proper watch on the trades. This is one of the problems of the 24 hour Forex market.

The balance of this system was down to about $3,500, so I called it a wash and deleted the account. I decided that since I can't keep an eye on this system all the time, that I would just go ahead and automate it.

So I stole the indicator code from the systems EA and integrated it into my own code-base. I feel much more comfortable with my own code-base, which I why I didn't simply keep trying to use this systems existing EA.

The result is the danw2 - weekly breakout v2 EA (code not available). Back-testing revealed profit on EURUSD, GBPUSD, and AUDUSD, which is a rarity. Here I'm running it on EURUSD, GBPJPY, and AUDUSD.

Already I can see problems would have developed on GBPJPY last week had it been running. See how the price bounced back and forth between the yellow lines on GBPJPY? This would have activated and stopped out several trades. It may be better to switch this system to the 4 hour time frame, but we'll see how it works out over a few weeks.

Basically, this system has been a failure, due to human error and laziness. I still think there is great potential in it. This is interesting to me, because it highlights the importance not so much of Forex trading systems, but the skill of the trader in the timing of placing trades and keeping a watch on them.


Daily Close Breakout 

 Last week, $812 profit, balance up to $5693. This week it went nuts, lost $1140, balance currently at $4553.

I have several other breakout robots, and they seem to do better with smaller stop-losses. This robot had a spectacular return last week, so I'll let it run another week or so before adjusting the stop-loss.


Open Close Bot H4

Lastly, my old H4 trend following robot. After several weeks of losses, it had a few good weeks, recouping all lost profit, before suffering a small loss of $39 last week, and getting absolutely cut to shreds again this week. This means choppy trading conditions! This is the only reason I really keep this robot alive - as another indicator of market conditions.

This robot needs several 4 hour candles to go in the same direction above or below the 10 SMA. This didn't happen this week.

This week it was down $587. Balance $4820.

EURUSD, 2012-03-22

The market has been a little wild this week, and I haven't been inspired to make many trades, certainly not on my live account. I've noticed the big jumps in the EURUSD and wondering what was happening. 

Now that the week is nearly over, it's fairly clear what was happening with EURUSD - we have this beautiful head and shoulders pattern. 

(60 minute EURUSD chart with head & shoulders pattern highlighted)

If prices can close beneath the blue line, they should keep going down. But as usual, it's Forex, so anything could happen. 

Robot Results 2012-03-11 to 2012-03-17

Alrighty, time for more robot results from 5 of the EA's I'm demo testing.

The attached results charts are smaller this week because of computer problems. My PC will no longer connect using remote desktop to the remote PC that I run most of my demos on - but my Macbook (which has a much smaller screen) will.


Inside Bar

This robot looks for inside bars and places pending orders above and below. 

Last week made 1 trade and $34.79 profit. This week I added AUDUSD. Three trades, all of reasonable size, but only $40 profit. Current balance is $4642. It's low because orignally this robot ran on the 1 hour chart and took a few big losses. 

A few more signals than last week though. There's also one trade open, currently up $108. 

Pending Reversal Candle

This robot looks for higher highs but lower closes (price exhaustion), and attempts to trade reversals. I added AUDUSD this week as well. It performed terribly on back-tests.

Last week this robot made $66.91 profit. AUDUSD didn't do us any favors this time, starting out with two big losses. It redeemed itself a little later in the week with a $330 win. The EURUSD profits are small for some reason. I'm sure there's something off with this code, but I can't quite figure it out. 

The balance on this account is currently $5341, with one open order currently at +$40. 

4 Hour Breakout Box

I'm trading this system from Forex Factory manually, using large lot sizes, while the EA trades smaller lot sizes and also closes trades and sets trailing stops. 

It was up $2669 two weeks ago. Last week the market dodged up and down early in the week, and $2000 of that profit was gone. This week the market did the same thing. I was a little more on top of re-adding orders once they were wiped out though, which made us a little profit (the EA is still closing orders too early a lot of times)

I got frustrated and placed some desperation trades on EURUSD and GBPJPY which both ended up being losers. Current account balance is $4788 with one order open, -$626.

I still have hope for this system, and it's easy enough to trade manually. There's some things which could be improved. First, sticking to the rules. Second, making sure I always re-add orders once they're stopped out. Third, probably disabling the EA and checking profit levels manually. Lastly, using the weekly close box only as a guide, and placing orders above/below several candles worth of H4 candles if they seem to be consolidating instead of just relying on the box. This invalidates point one, "sticking to the rules", but common sense should play a part too.

It's easy to see how this system could be successful, and it's frustrating to miss the trades which would have made it so because I believe the market can or can't do something and don't follow the rules.


Daily Close Breakout

I added this system last week. It made $812. This week was not so spectacular, and I had almost given up hope (it also performed poorly in back-testing). But it's down only $81.32. Current balance $5693, one order open, -$53.


Open Close Bot H4

This is a simple trend following robot. It goes through stretches of winning and losing. I've been running it for a long time and keep it mostly to gauge how strong trends are during the week. If the robot wins and makes profit, the market is trending strongly, or at least moves in one direction long enough that this robot can detect the trend and place winning trades.

The previous two weeks were very good, and this robot made back all that it had lost over the previous month. The week just gone wasn't so great, and it's down $39.54, with no current open orders.

---

Overall, a pretty disappointing week for these robots. These results are pretty typical. Actually, they're not as bad as they usually are. Although some of these robots had successful results in the previous weeks, success seldom lasts.

Interestingly, some other robots which I am demo testing did OK this last week. Previously they were sucking, but small settings changes seems to have helped. However, I'm not going to know for sure for another few weeks as the market continues putting them through their paces.

Testing is a very slow process.

EURUSD

Not too much movement in EURUSD last week. This currency pair opened and closed within last weeks opening and closing range, with a brief test of the 1.30 level again.

(Weekly EURUSD chart showing last weeks candle sitting nearly inside the range of the previous weeks)

Price goes down, price goes up, down, messes around, then shoots up on Friday but only slightly exceeds the high posted earlier in the week. Not a good week for most of my automated systems, and not too great for manual trading either.

(4 hour EURUSD with blue line indicating previous weeks closing price. The vertical dotted lines separate weeks.)

I noticed the price did seem to be continuing downwards, at least until Friday, and I had several attempts at drawing trend lines, all miserable, before finally getting this:

(60 minute EURUSD chart with trend line in orange. Blue line is the 1.30 level)

It appears price has only slightly exceeded my trend line. Will it continue going up?

On this next chart with simple moving averages plotted, we can see that price hasn't even reached the 100 SMA on the daily again.

(Daily EURUSD chart with 100 & 200 period simple moving average)

But on the Ichimoku Kinko Hyo, we had a little cloud penetration then a sharp move up.

(Daily EURUSD chart with Ichimoku Kinko Hyo)

So once again, it's really anyone's guess, but price does seem to be trending downwards. However, it has to penetrate the stubborn 1.30 level to keep moving downwards. Additionally, several other currency pairs also traded higher on Friday (AUDUSD and GBPUSD), so we'll keep an eye on them for hints as to what EURUSD price action may do. 

As far as my breakout system that I've been trading manually on a demo account goes, it was rather shameful last week. 4 trades, 2 break evens, and 2 losses, totaling $80.60. I did have some success with other currencies pairs so my total loss for this account was only $30, but my EURUSD trades have been pretty horrible. 

See from the chart below where Fridays trades went wrong. I was faked out in a sell, then my BUY trade broke even. Pretty nasty. 

(30 min EURUSD chart. Yellow box represents consolidation. Pending orders were places at the top and bottom of the small blue box. Dotted lines separate days).  

Quite often lately I've found myself thinking I should ditch the Euro. We'll see how it goes next week.

EURUSD

Not too much movement in EURUSD so far this week, just a lot of chop which has been playing havoc with the breakout robots and systems I'm testing.

(60 minute EURUSD chart with 100 period SMA high & low plotted.)

The yellow box highlights the first 4 hour candle created after the weekly open. Price has passed into or through this box 5 times already.

Getting frustrated with my breakout strategies, I realized price is actually still moving - in this case, downwards, as indicated by the orange trend-line. It only connects two points at the moment but we'll see how it holds up.

The 1.30 level is very close, and price is still just resting on the top of the Ichimoku Kinko Hyo cloud on the daily chart (not shown). 

GBP is trying to head off in the other direction, so it's not providing confirmation of a move lower how I'd like it to. We'll just hold tight and keep watching. 

danw2 bb exceeder + high-low + pending EA

This robot with the long name, available here, uses my latest code base, and offers reasonable back-test results.

(Back test from 60 minute EURUSD chart, 1-1-2011 to 1-1-2012)

Almost a 45% return over a year, winning 70% of all trades, with a nice slope to the equity curve.

With optimization, it even generates some profit on GBPUSD:

(Back test from 60 minute GBPUSD chart, 1-1-2011 to 1-1-2012)

Not quite as impressive as EURUSD, but a 12% return and 76% win rate.


It hasn't been demo tested yet, so may yet fail. If you download the robot, it's set to run on EURUSD on the 60 minute chart. You'll have to adjust the settings for other currency pairs. The only other pair I back tested on was GBPUSD.

It works by waiting for candles to exceed the upper or lower bands of the Bollinger Band indicator. Then it places pending orders above or below that candle, waiting for a reversal.

This is the first of my modern robots that back tests well which I am releasing for free. Use at your own risk.

EURUSD

Thursday of last week saw the EURUSD climb rapidly and stall, before falling all the way back on Friday.

Above is a 60 minute EURUSD chart with 100 period high and low prices of a simple moving average plotted. I love the way the price seems to accelerate through these lines!

I drew a bullish channel on the daily chart at the beginning of last week, but the price broke out of both sides of it, and now with this latest bearish move, the channel is well and truly useless.

Below is a daily chart of EURUSD with 100 and 200 period simple moving averages:

The price rose Thursday, hit the 100 SMA, and fell. Most of the other time-frames also show price action underneath the 100 SMA which means in theory there's nowhere for it to go but down. 

But, beware. Traders looking to buy EURUSD will use this opportunity to do so. We have the 1.3000 level to penetrate first before we can be sure of continuing bearish moves. 

Looking at the daily EURUSD chart with Ichimoku Kinko Hyo plotted we can see price is just resting on the cloud - it would need to penetrate this as well if it is to continue falling: 

So, technically everything is pointing towards bearish price movement. Even this pretty lady with the excellent charts agrees. 

If there's a big fight between the bulls and the bears, then EURUSD will consolidate and whipsaw around like a wild caged animal for a while. That won't be fun. 

As always, I'll look for consolidation patterns. Here is a 5 minute EURUSD chart with 100 period Bollinger Bands plotted. It's already formed a nice consolidation pattern, which I have missed to the down-side. I will consider setting a BUY-STOP to the top-side, though.

dan mtf adx robot

This robot from 7-5-2010 is a mult time-frame ADX robot. It checks the value of ADX(21) on 5, 15, 30, 60, and 240 minute time frames, then reads RSI(8) on the same time frames to check for over-brought/over-sold conditions, as well as averaging the results of the stochastic indicator from the same time-frames for over-brought/over-sold technical information.

I was big on over-brought/over-sold back then.

It didn't back-test well. So I checked over the code, found and fixed some (hilarious) bugs, adjusted take-profit, stop-loss, trailing-stop, ADX and RSI values, and tried again.

This result is actually from about the 10th attempt (back-test on 5 minute EURUSD chart, 1-1-2011 to 1-1-2012). RSI and ADX were set to 20. 

$1010 loss, 38% win rate. Not great at all.

One of my multi time frame ADX robots worked fairly well for a while on a demo test, but I don't think it was this one. I think there are several robots with the same name, as they all operated on the same general principal.

Robot Results March 5th-10th

This is quite fun, reviewing robot results running on demo charts from week to week. Last weeks post and results are here.

The code for these robots is not available at present.

Generally my robots don't do very well. Out of the 30 or so I've been testing recently, they showed some success weeks ago, went through a period of poor results, and then in the last two weeks many have had fairly good results again. I would even call them optimistic, except I know their performance is just a result of market conditions, and once the market changes again they'll give their profits back.

So I've been reviewing 3 robots - Inside Bar, Pending Reversal Candle, and 4 Hour Break-Out Box. I'm going to add two more just for comparison - Daily Close Breakout, and Open-Close Bot H4.

Inside Bar

Last week this robot made $34.79 profit over 1 trade. This week it's down $314 in two lost trades. Interestingly, it only detected one inside bar. It placed trades above and below and got whipsawed out of both of them. I gave the code a quick look over looking for bugs, but I didn't see anything obvious. I might try running this on another currency pair next week.

Pending Reversal Candle

This robot places a pending trade only in the direction of an arrow, waiting for a reversal. Only two active trades this week, for $66.91 profit. Considering the leverage this robot uses, this is really quite low. I did close both trades manually, however, and the second one would have made a little more profit if I had left it. The first one could have been stopped out for a loss if I hadn't closed it, though. 

Also, it draws a red box around the last few bars, and actually places the trade above this box. The box drawing seems to be delayed and I'm not sure it's placing as many trades as it should. I'll need to review the code for this robot also. 

4 Hour Break-Out Box

This is a system and EA available from Forex Factory here. I'm allowing the EA to place trades (using 0.01 lot size), but am also trading it manually (using a lot size of 1). This week I added AUDUSD to the currency pairs I'll trade with this robot.

Last week it was up 53% with $2,669 profit. This week it didn't do so well, giving back almost all of it's profit. 

The problem was the market whipsawed like a crazy witch around the weekly open line, which this system uses to draw it's break-out box. Also, the 1st 4 hour candles of the week were very small so the breakout box was small.

This system could have avoided it's losses had break-even been set early after trade activation. I believe the EA does this, but not soon enough, apparently. Also, I wasn't consistent in setting orders. It could have recovered a little of it's losses from EURUSD on Friday had I set an additional pending SELL-STOP order, but by then I had pretty much given up trying and didn't think EURUSD would be able to descend that far.

Also, AUDUSD took off shortly after the week open and I completely missed an opportunity to capture any pips on that pair.

I still have hope for this system. Since I'm now trading 3 pairs, I should reduce the lot size and also monitor it a little more closely. I had pretty much neglected it this week. Or perhaps it's better to focus only on one pair?


Daily Close Break-Out

 

This EA performed great in back-tests until I discovered a serious bug which meant it wasn't actually working how I thought it was at all. The bug is fixed, but now the back-test results are horrible!

I'm fully expecting this robot to fail, but was very surprised by this weeks results. Up $812. 

The complete list of orders won't fit on the screen, but you get the general idea.

Open-Close Box H4

Once I coded a very simple robot which opened a BUY trade if the current bar opened higher than the closing price of the bar before it. It had a 5 pip stop-loss, and made over $100,000,000 in a 2-year back-test on EURUSD.

It was shortly after that I discovered just how screwed up my back-test history data was. Quite sad. I was very excited for a while.

This robot is based on that design, but it adds a moving average and runs on the H4 time-frame. When I first started back-testing it some months ago, it performed great for a few weeks, then started sucking. The demo account expired and I created a new one, and it's been performing poorly ever since - until last week.

It was down to about $4,000, but in the last two weeks has made about $1,800. There has been some chop  on the EURUSD, but apparently prices have trended enough for this robot to catch some good trades.

I closed a couple of trades manually in this last week, when I noticed the price beginning to lose momentum.