The Captain: an always-in auto-trader by SteveHopwood

This ea has the rare accolade of making it onto my demo trading machine. Given that I code about 5 ea's a week, you can imagine how rare this is.

It has the even rarer accolade of surviving a few days on my demo.

Early this week, demoing on a D1 chart I was finding this: early gains were wiped out during the following candle, leaving the account down despite all that lovely early profit. This is hardly surprising; the idea that a second candle will follow the direction of the first candle just because that candle went in one direction, is infantile.

So, I set up mptm to close all trades at a profit of 1% of balance. Since then, I have made back all the losses and the balance is about 1% in profit.

So, this ea survives both my demo and my interest into next week.

Nice one Cap. Looks like your idea isn't total bollocks, unlike the vast majority of the stuff that people send me in pm's.

I am leaving the ea on during nfp tomorrow, just to see what happens; if the result is not good, I will not blame the strategy. Only an idiot would trade it live tomorrow.

Looking forward to some entertainment tomorrow.

Anybody wanna discuss this chart? It's price movement inside daily candles.

I do not have an analytical mind (aka not too bright), so I'm not sure if my conclusions are viable or not. But, if the plan is to trade in 'continuation mode,' then, obviously, everything's good when green follows green and red follows red. Now, the pair in the chart, the GBPUSD, has a habit of producing good-sized wicks on its daily candles. So, even when green is followed by red, and vice versa, there's usually enough price movement in the continuation direction to make for a profitable trade, or at least a 15-30p BE stop.

Granted, there are days when the wick isn't big enough, so stops would have to be in place, but they are fairly uncommon. The second chart is typical of the GU. Big, yummy wicks, day after day.

I was looking at running a separate account where trades could be placed in the opposite direction (short after a green candle, long after red) to take advantage of the 'wick pips.' I think there's something to be gained here, but I haven't found the time to put it together yet. Or I could be dreaming again.
Anyone care to take a stab at it?

[quote=SteveHopwood;4597690]

So, I set up mptm to close all trades at a profit of 1% of balance. Since then, I have made back all the losses and the balance is about 1% in profit.

Hi Steve,

in one pair or in multiple pairs?

Thanks

Cap, I owe you an apology.

I thought this idea was faintly ludicrous when you sent me your pm, and coded it only because it required no effort on my part. I am not sure why I even bothered to put it on a demo except that bBacktesting showed modest profits on H4/D1 on a couple of pairs, and LOJ (Law of the Jungle) had killed off a couple of other bots so I had some spare capacity.

Early days I know, but the start made by this bot is promising enough for me to start taking it seriously. I am about to post a link in TB's thread to make sure The Brains turn up here for a look.

Pass your eyes over the attached, folks. I restricted trading to the majors to keep the dd at bay:

 

1. Monday was a losing day, yet the trades had been in good profit at one stage. What happened was: the profits dwindled through the hours leading up to the formation of the new candle; the trades gradually went into loss; they were closed at the formation of the new D1 candle at a loss.
2. Tuesday: modest profits. I noticed that these were much higher earlier in the day. I see I closed the GU sell manually to keep the profit; the remaining trades dwindled etc
3. Wednesday: all this persuaded me to set up mptm to take a global closure at 1% upl; you can see the times at which this target was hit. Had the target not been hit, then TC would have carried on trading as normal at the start of the new candle.
4. Thursday: 1% profit target hit.
5. Friday: dd has been kept pleasantly under control, so I doubled the lot size. Clearly, the more pairs you trade, the lower needs to be the lot size; it looks as though 0.1 per $1000 is ok on my IBFX Aussie 1:100 account.

I think we should start looking at the idea of balanced trading that emerged recently in TB's thread. This was a new concept to me, so for others similarly ignorant here is a brief explanation:

 

1. a seismic event that affects one nations currency will have a knock on effect on all pairs that include that currency. The recent earthquake/tsunami that hit Japan perversely strengthened the Yen dramatically. Those of use with open TB buy xxxJPY trades saw them hurtle dramatically into drawdown - conversely those lucky enough to be on the sell xxxJPY side saw the profits accrue quickly.
2. so, anyone who, say, had a buy GBPJPY and a sell USDJPY would have seen mostly a neutral effect. Closing the UJ trade on the news that world banks were going to act to weaken the Yen and hanging on to the GJ until it broke even would have netted a nice profit overall.
3. 'balanced' trading therefore means hedging a currency against itself, so:
   1. buy usdjpy, also buy, say, gpbusd: this neutralises the effect of a major move by the dollar.
   2. having bought gu, sell gj or buy eurgbp: this neutralises the effect of a major move by the pound.
   3. etc

My thoughts for TC's future development run long these lines:

 

1. at the start of the new candle, take trades as normal with two or three pairs. I will take GU and EJ as my 'trigger' pairs.
2. supposing GU is a buy and EJ is a sell, at 5 minutes into the new candle:
   1. each instance of the bot will look at its own pair and send a trade if this can balance one of the triggers, or example:
      1. sell GBPCHF OR BUY UC
      2. buy EURCAD

I am not sure if all this is necessary, and I am not going to do anything about it until I have seen the bot at work for at least another week. It does give us something to think about.

Pinhead, I will code a TradeOppositeDirection in time for Monday, so you can explore the potential of trading against the previous candle's direction.

" Do not go long with a green candle below the 6 SMA"

do you mean the close or open or high or low of the candle should be below sma ?

Hi Steve,

These types of system are quite well known & I believe originate from statistical work done on how often red/green candles & combinations come up. Somewhere there is a body of work done on this.

Another slant is to go long if you get two green candles following each other and vice versa - go short if you get two red candles following each other.

There are also more complex combinations

The image attached shows that often the two green/two red can work quite well - not sure about long term however.

Enjoy ....

Cheers macman.

Im trying to test on demo and each time it tries to enter a trade I get the message order failed with error (130) invalid stops.. I have the stops and TP set at 20???

Great catch macman! You've done it again!

Ditto.

Hi

I tested the auto trader during the NFP news release in M5 TF for three pairs GU EU and UC, the result is $280 trading with 0.1 lot .
The MT4 platform crashed twice, I bellieve (without being an expert) because of the error message that gives when multiple pair are traded. With the release of the news it gave also an opposite direction trade that was later ignored by the EA and i had to closed manualy.

EU and GU gave better results than UC. I had to stop testing as I couldn't open more trades because my demo has low leverage and I run out of fuel.

Attached are the trades

Those of you testing this system, what settings are you using for SL, TP and order size? or are you using it straight out of the box?

OK, on my live trading account, current balance $202:

 

1. Lot size: 5 lots
2. tp: 500
3. sl: 5

As you can see, my cunning r:r of 1:100 ensures my account will grow magnificently in all market conditions. I will be rich by this time next week.

I tried low TF with the lowest SL TP only during the NFP was out of the box.

I wouldn't tell somebody this I'd tell the truth.......I don't use a SL

Cheers. What surprises me most is the great difference in back testing between different pairs using the same settings. One pair the graph will be bottom left to top right, and the next pair will be top left to bottom right.

However this is coming from an intaday trader who has only ever traded 4 pairs..

As macman brought this up, I like to add my 2 cents....

I normally trade H1 and H4 tf and used the probability distribution of said candle sequences with some success in the past.... The attached indicator visualizes this and allows anybody to play with different candle counts on their preferred tf.

I found that 5 same colour candles on H1 gives a fairly high probability of reversal on most currencies...

So something along the lines, open order on 2nd same colour candle, dont add after 3,4 or 5 (user selectable) might be an idea.....

Steve - could you please have a look at my "Sushi Roll" post in the newbie thread..... Appreciated

For those who want to play, I have added TradeOppositeOfTrigger; this tells TC to trade in the 'wrong' direction.

Have fun.

I have seen it and will code it some time.

Great news guys. I just had a bath-lol.

This is merely The Beast with a different trade trigger, so all I have to do to code it is change the Sixths trigger to that of an x-candle reversal trigger.

I have bin a-learnin'. A contributor sent me some fantastic information about using Rsi last week. I am not at liberty to post this because it is copyright by the authour and not freely available.

I am able to make use of what I have learned; there is stuff in the pipeline that could make TB redundant. We shall see.

In the meantime, have a look at the pictures. The first one shows the eu fall on Friday - the vertical line is the candle just before the fall began. On the Rsi window is an SMA (3) of the Rsi - if you do not know how to drag a MA onto an indi window, use the attached template.

What I have learned is this: the best use of Rsi is as a momentum indicator to warn of an impending price movement. You will see that every decent price movement was preceeded by a cross of the Rsi line over the Sma.

The second picture is an up-close-and-personal view of the start of the fall; the rsi/sma cross is clearly visible.

So:

 

1. first we have a change in momentum signalled by the rsi/sma cross
2. then the price follows on the chart (I made a fair few demo pips scalping this last week, and the ea is almost complete).

I can create an ea that retains the best of TB with box trading, Recovery, Volatility et al and simply change the setup/trigger to:

 

1. setup: start looking for a reversal after x candles have moved consistently in one direction
2. trigger: rsi crosses sma in the appropriate direction

Keep the ideas flowing guys. You suggest them and I will code them. We all know that most will end in the bin, but we only need to score a success every now and again to keep well ahead of the trading game.