Steve Hopwood's Epsilon Megastar fib auto-trading robot

I decided to change the color on my Draw Daily Fibs approx 15 mins after the 2 Sell trades opened

Upon the EA reseting the Daily Fibs color the Yellow Breach 50% reset 35 pips lower, approx 1.2290 - 1.2255

is this correct ?

Yes. The robot picked up on market conditions as they were when it was fired up. It is not coded to take any notice of existing trades.

I am already thinking about ways to make the robot pick up where it left off - a complicated process in this case.

When posted for wider consumption, the robot will be issued with a warning not to to mess around with it once trades are opened. In large red letters. Using the term 'idiot' frequently. Usual polite, nurturing sort of thing.

Resetting Compiling Changing EA's after trades are open is my specialty

the ultimate EA test !

Per the video clips, isn't the breech of the 50% fib supposed to be if price 'closed' past? Not just a price breech of the line.

A pic showing the Breach Pivot move, and showing the market price pulling up strong to the Breach Pivot before heading south

1: First Breach Pivot Position

2: Strong Market movement back to Breach Pivot, failure to break it

3: Second Breach Pivot Position after I reset EA

I was thinking that there probably needs to be a minimum difference between the old and new pivot. To automate it conceptually, there might be an indicator-based (such as ATR multiple) method to set a minimum distance between pivots, as you suggest so that the bot will keep looking for previous fibs until it finds one that meets this minimum.

I noticed in manual testing the system that there are many cases where a previous fib is "valid" but it is so close to the the current fib that one would have 2 entries rather close together. I'm not sure that is a good idea if automated without user intervention because it can quickly increase the lots on a single pair if the market moves for an extended period with such small fib moves. A minimum pivot distance would allow the user to control this.

Also, note that Taz mentions in his post that the strategy calls for a close of the candle after the pivot breech. Is the bot currently coded to do that? A close of the candle might alleviate the concern about the bid vs. ask issue. Although, it still might make sense to have a user-configurable "pip add-on" for long trades, as well as one for short trades, particularly if the candle close is not sufficient to ensure good entires.

To show you the trade so far

1: First Breach Pivot position

2: Strong pull back to Pivot

3: Second Breach Pivot position after I reset the EA

3: Third Trade opened as price closes below Breach Pivot

4: Take Profit, price comes 2 pips away then pull back up ! lol

Since the strategy also is intended to take into consideration candle patterns, this might be something to add to the bot at a later date. There are some open-source bots that check candle patterns (at least I think I've seen them before somewhere) whhich might be helpful in coding this into the bot.

In my manual testing I have also noticed that when price is consolidating with shorter candles and/or inner/outer bars, price can breech the pivot and the candle may close to trigger a trade on the next candle, but price may end up breaking in the unintended direction as it moves out of consolidation. If manually trading, one can look for these conditions, and choose not to take such a trade, but how might the bot accomplish this? Or does it matter? The stop loss might be pretty tight anyway and if the bot has a high win rate overall, perhaps "filtering" for consolidation does not matter to the bot's overall profitability.

No 2 is one of those 'waves' that Simone describes in her video. I am already in short on GU; if the intraday pivot rejects the price, then I shall add another trade on the break of the breech pivot.

Se, we need:

 

1. minimum distance between old and new pivot
2. a user-definable pips-add-on to the distance between the pivot and the opening of the new candle (and yes, guys, the robot waits for the start of a new candle before checking if a trade is possible). Rather than add another input, I will try adding the spread into the calculation to see if this makes any difference.
3. a 'pips tolerance' to add/subtract from the take profit to alleviate the situation Burton described, where the price gets to within a sniff of tp then retraces. When using fibs to generate stops and tp's manually, I have always done this.

Just taken my first live manually-traded pips using this system.

Granted, they are EU pips and a child would say sell EU right now, but still manually-traded live pips.

That gu 'wave' rejection I was waiting for has happened, although it took a retrace to the I Long for it to happen.

An interesting problem has manifested itself this morning. The EU is now so low that the only way to find a previous low for the breech fig, is to go to a monthly chart to find one - back on 1st Nov 2005.

This means the robot cannot draw a breech fib as it is only programmed to look at 15M candles and there are only so many that a chart can hold.

This means programming the previous low finding routine so that it will gradually move to higher tf's if it cannot find a low on the current one.

I thought this was going to be an easy bot to code. It is turning into a nightmare.

Hey-ho. Back to the drawing board.

Again.

I have done some updating. Some of this was complicated, so there may well be bugs lurking in the undergrowth.

 

1. the robot should pick up where it left off after a reinitialisation, for example after loading the latest version. Every time it adjusts a fib, it stores the values of the hi-lo in global variables which it reads when restarting. Youcan view these by pressing the F3 key - they all begin with 'Simone'. Any that do not begin with 'Simone' were placed there by something else, so leave them alone. If you want to force to bot to restart from scratch, simply highlight the relevant gv's and delete them. This code is where bugs are most likely to appear.
2. I have put the fractals-finding code inside a loop that allows the robot to look at higher tf's when the current one does not have enough bars to find a breech hi-lo. A distant hi-lo will look a little strange on the chart.
3. Added the spread to the pivot-break calculation. If this proves to be insufficient, I can add a few pips more. I just looked and the demo has taken fresh tranche of trades; adding the spread appears to be sufficient. We shall see.
4. Added a pips tolerance of 5(0) pips to the take profit calculation to cater for those near-misses-by-a-couple-of-pips. This can be a user input if people want.

The post trade-send breech fib movement is definitely wrong. I need to think about it. Probably a test that the new pivot is the correct side of an open trade's open price is needed.

Thank you Steve for the continued development of Simone.

Using the previous version the bot has generated 1.7% trading 12 pairs. I could live with results like that every 24 hours

My EM demo on H4 with mptm is holding onto a profit. The trades are infrequent, and they stay open for days, but so far so good.

Here is the latest.

 

1. I am trying the brutalist approach with the post-trade breech fib - deleteing everything to do with it in the hope that this will force the bot to recalculate it properly. If it doesn't, you will hear the scream whereever you live.
2. Sorted out some bugs in the tp adjustment.
3. I have added code to make the robot change trading direction and redraw thebreech when the market flips either side of intraday pivot. It remains to be seen what effect this has, and we still have to decide what to do about trades left open in the wrong direction, so say nothing of the fact that you poor blighters in the US are not allowed to hedge. Sorting this out is for the future.

For now, I just want to get the blasted breech displaying properly after a trade.

Steve,

Your dedication is out of this world. Also, After you get this figured out can you solve british oil company's (BP) bloody oil spill onto our yank beach's

Steve,
Don't listen to him. He doesn't know a good thing if he sees it. Heard of 'Beer on tap' but 'OIL ON TAP'?

Has anyone noticed the bot taking up a lot of PC resources? I'm finding my VPS at 100% CPU usage and if I close down the MT4 platform running the bot, CPU usage returns to normal. When I restart the platform, it works for awhile and eventually ramps up to 100% CPU usage. This began to occur only after adding the Simone bot.

BTW, I'm using the version of the bot from yesterday before the last two version updates.

Same here