Steve Hopwood's Fib Target-level retracement auto-trader: compliments Simone

The most recent version posted does this. Go back a few posts.

Actually, I feel a little irritated here.

I really don't care what your whole report is.

You appear to have ignored my bugfixed version.

You appear to have ignored the question I asked a few posts ago.

And, no, I will not repeat it. If you couldn't be arsed to read it first time around, I see no reason to repeat it.

One of the things I really, really, reallyhate is when people prove conclusively that they cannot be bothered to read my posts in response to an idea they have themselves posited. As though I have nothing better to do than repeat myself.

Actually, I am not desperately impressed when people prove conclusively that they cannot even be bothered to read my posts at all. So, imagine my fury when they ignore posts aimed specifically at them. On second thoughts, better not to.

Steve-sorry for the delay in responding, I had to sleep a few hours.

I'm sorry also about any mix up with the bugfix version. I did download and replace the version in the post above before asking you to post it. It just still wasn't working for me, thats why I asked. I've done so again, and I'm still just getting buys. Not that thats your problem, but just to explain that I am definitely 100% not ignoring anything you write or post-to the contrary I am reading and re-reading everything you write.

As far as the question you asked me about the stopp-loss, I thought I did answer you in my previous post, unless I misunderstood your question.

I took your question to mean should I change the SL function to stop only when it hits the average line, instead of a numbered stop. If that was your true meaning, then my thoughts were that sounds good as long as the max distance function works so you don't get any very large loosers, and there is a min SL function to prevent getting stopped out if a position is opened right near the average line.

If you meant something other-then again humble apologies. I really am reading carefully and not taking anything for granted. I think its an amazing bot already and you've done an amazing job.

Anyway, thanks for all the hard work. It is always appreciated by myself and I'm sure all concerned.

This version has a check that the stop loss is on the opposite side of the trend line by at least 5 pips. You can have this minimum distance as an input if you like.

Also attached is a M15 GJ backtest with some sell trades.

This one is looking promising, so I shall add it to post 1 over the weekend, so people can find it easily.

Excellent Steve-thanks for that!

I just wanted to mention some results for backtests on the retrace part of this EA. I saw no reason to ignore it since you took the trouble to include it and everything.

For the sake of any readers at all confused: The retrace part of course was designed for the H4, and should only be tested and traded for that TF.

The breakout part uses and yields profits on the 15 Min. So on no account will they work for the best if both are enabled on one chart. You have two different bots in one package here.

In short it seems the JPY pairs love the retrace, the UJ in particular. That may have to do with the regulation and intervention policies of the BOJ. I know at least one well renown programmer/trader has built a whole system around that.

Anyway-UJ yielded 32% over three years with only an 11.6% Max DD. Better return than the banks are giving, thats for sure. So point being there is potential there as well.

-Cheers

(I might have to eat my words about retrace not giving good results on 15 min. It is testing nicely on several pairs. I'll leave it to folks to do their own research, just wanted to point out its worth looking into)

I have uploaded Monster Muncher to post 1 and redesigned the post. Future updates will be uploaded to post 1 so they are easily found.

Here is something for you guys to play with. Asjeff sent me a pdf and there is some thought-provoking stuff in it.

Amongst other things, the pdf describes how to use Atr to detect when a market is trending or ranging. I cannot post the pdf here as it is a commercial product. I have provided a precis of how I am using information gleaned from the pdf - attached.

The script returns the nature of the market. I turned this into an EA, so you can play with both.

Also attached is V1 of ChandelierStops - the pdf featires its use in exiting trades. I found this by doing a quick search of FF. Anybody got a later version?

Basically, ChandelierStops is an indi that moves upwards/downwards in relation to new highs/lows and market volatility. For example, in a rising market, the indi follows the new highs at the distance you select in the inputs. If volatility increases, CS moves closer to the market. This can be used to calculate rising stop losses, locking in profits/cutting down on losses.

I am not sure if any of this has an application to these robots. It might be a help in manual trading and for those of us trying to devise and develop our individual trading methods.

Fredella has just pm'd me to point out one of my usual daft copy/paste in both the script and the ea, so the fix needs downloading.

One day, I will learn to check my copy/pasting with a modicum of care.

Hi Steve et al,

Hope everyone is having a pleasant weekend. About 95 degrees in the shade here in Florida.

Anywho, after looking more deeply into the retrace results on the 15 min TF-boy was I wrong when I said the breakout and retrace can't work together. (As far as backtests are telling me, so all still theoretical at this point, of course.)

EJ showed the best, pulling in 3500+ pips in a six month testing period, with both retrace and breakout activated on 15 Min.
Jan 3-July 10
SL: 1200
TP: 1800
RSI: 71, 29
Sleep: 4 min

If you add up what the two sides made when traded separately at the same settings during the same period, it only came to 2600 pips. This has also occured on several other pairs. So something seems to be going on there where the order placement patterns become advantageous when both are working together.

When I said earlier this wouldn't be the case, I simply spoke too soon. That was based on my feelings that the retrace wouldn't work that well due to earlier analysis, not the evidence of the backtests. So, I happily eat my words. Retrace backtests profitably both on its own and also with breakout on 15 min.

----------------------

I just also wanted to express a (belated) thanks to the immensely talented Mister Hopwood for coding this. It was very exciting to me to see these ideas get realized in the form of an EA. Once I got the bot, my head got a bit lost in the data, and I neglected that-So a big Thank You Steve!

----------------------

If you have any interest Steve, I've developed a logic upgrade for the breakout method which could be applied to the 5 min charts.

After studying the breakout results on the 5 min I realized that most of the losers were coming late in the week-Wed-Fri. This makes sense because the initial projections the 5 min are making are only based on 1/3 the time that the 15 min are based on. So by Wed they have become irrelevant, and its just like placing random trades based on RSI OB OS.

The solution I developed goes like this:

1) The projections update themselves every 4 hours as long as a trade has been closed in either a profit or a loss.

2) If no trade has been taken based on the last projections, no update will occur in the next 4hr period.

3) If the EA is currently in trade, no update will occur in the next 4 hr period.

4) A maximum of two 4 hour periods can be skipped. On the third period, an update occurs weather or not the bot has made a trade based on the last projections. Although as a variable input: Skip Max Periods, it could be tweaked.

5) This version of the bot couldn't have the retrace feature. It would stand on its own as an M5 scalper. All the other features and inputs would be just the same as the Pip Muncher.

The idea here is to keep the data fresh so that its constantly relevant to current price, but also give the price enough time to form a deviation from the defined trend and take a trade.

I took a random period on the EJ 5 min chart and got 4/5 winners at 50 TP to 50 SL basis, over half a week of testing. So maybe there is a 300 pip/wk potential there, which is a little better than double what I can get right now on my best backtests with current EA.

Anyway, I'm just putting this out there as an idea. I'm greatly appreciative of everything and look forward to doing some demo with the Pip Muncher this week.

Hehe. This is like keeping up with a puppy endowed with brains and experience.

Steve, thanks for your kind appreciation. The fact is, I can only code good ideas as they come in. Without the good ideas, I am useless. I am just the mechanic that bolts everything together. I am coming to love your posts.

My demo account from Monday should be fun. Loads of M15 charts set to trade both the breakout and the retrace. Fabulous.

Re the 5 minute thingy, I am not sure if it is possible to modify the bot to take care of this. Could you add this to your list of things to nag me about, please, and give the rest of us a few days to demo the existing robot to see what happens?

In the meantime, I will look at updating the code so that it takes care of the M5 stuff. Once this is done, it will then be possible to accommodate your imagination as fresh ideas occur. I forsee a monster trading robot emerging able to take care of all sorts of trading conditions.

Absolutely Steve, Monster Muncher needs some TLC this week. Its just better for me to get idea out so I don't have to think about it anymore. Helps me sleep.

-Cheers

My thanks to others who are running the backtests. Nice to have a quick report on the ea's being developed here.

I began a Forward Test of the FTLR 2 weeks ago. See post # 284 for the 1st weeks results and explanation. Results for 2nd week as follows.

// ------------------------------------------------

Daily
Closed Activity / 46 trades / all positive / total 2915 pips ($29.15)
Open Activity / 6 active trades / 4 negative, 2 positive / net -495 pips (-$4.95)


4-hour
Closed Activity / 65 trades / all positive / total 426 pips ($4.26)
Open Activity / 10 active trades / 6 negative, 4 positive / net -767 pips (-$7.67)


1-hour
Closed Activity / 115 trades / all positive / total 1556 pips ($15.56)
Open Activity / 11 active trades / all negative / total -1337 pips (-$13.37)

// ------------------------------------------------

This is from 17 pairs with NO monitoring at all. All time frames still positive. Daily with the best equity position. Will continue letting them run for now.

Pip Parade

Nice one Pip Parade. Thanks.

Fib retrace still at 100% success rate - 23 trades average win 6.5 pips per trade.
6 pairs on M15 - out of the box settings.

See earlier post for more details - http://www.forexfactory.com/showpost...&postcount=286

Excellent macman. Cheers

I have been watching a video made by a highly successful trader, talking about use of Atr for trend-spotting and for adaptive stops and tp's.

This is a commercial video, so I am not able to post the video here. I can share what I have gleaned from this, because in doing so I am merely disseminating my understanding of points made by a third party.

One of the strongest points made by the author is this: if a trading system is to remain successful in the mid to long term, then it must have exits that adapt to market conditions. 'Hard', pip/cash based exits will fail sooner rather than later, when market conditions change.

This makes sense. Simone and FTLR already have adaptive qualities; if market conditions change then the fibs will change with them.

I have added the adaptive qualities of Atr to more of Monster Muncher's functions. The update is in post 1, along with a description of the update in a new MM Updates section.

Have a play guys, and see what you think. The default settings are leading to more profitable backtests, but there are a mighty host of different combinations to play with.

40 consecutive winners, no losses. Negative float less than 3%, currently. Running 16 pairs on auto-pilot. IBFX demo with 10-cent pips.

Great results here also. Fabulous to read.

Can I ask people to say what time frame they are trading and what target level they are using when reporting results? It would be a help to the rest of us.

Time Frame=D1

Not sure what you mean by target level. Please explain.

Properties, as shown below.

Thanks. Target levels do not apply here as you are trading the multitrade thingy.

You appear to be trading pretty much as I originally intended; great to see this works so well. Please keep posting your results.

Cheers