Nanningbob V5: Steve Hopwood's auto-trading robot

Hedge funds by their nature will often hedge any negative forex exposure against other instruments which are not normally available to retail traders, like bonds etc.

From my experience/knowledge then hedge funds thend to make between 0.25% - 0.30% per day on average. Some retail traders may think this is a small amount. But its not when they are using $25M worth of tradeable assets.

Keep up the good work people.

Scoobs.

[quote=scooby-doo;3968643]...hedge funds thend to make between 0.25% - 0.30% per day on average. Some retail traders may think this is a small amount...quote]

It makes me laugh when I see my fellow traders complain 0.30% per day is not enough. I would be more than happy with such rate of return even on $1K. Do the maths 0.30% x 365 = 109.50%.

Kind Regards
P.S. Thanks MAX for correcting me. I counted the weekends and public holidays also and forgot about the compounding factor. I fear due to the success of the manual system in August people come to expect such high returns from Nanningbob system. However, Bob says that August is the best month for his method, therefore, the following months would be a better representation of the ability of this system.

Me too. They should feel chuffed to have left the 95%. I sure so.

not compounded....

0.3% x 250 trading days = 110% a year (compounded), 840% for 3 years (compounded)

This may seem like an odd request - but during my backtests of the Trend and SAR EA I'm only seeing orders closed from takeprofit or emergency stop loss - (or closed by the end of the backtest, where all orders are automatically realised) no single trade has been CLOSED.

I can't actually see anything in the EA code which points to why the trade closure code wouldn't work, so two questions
1/ one has anyone demoing NB Trend and SAR v1.00 or v1.01 seen a trade close - (not by take profit or stoploss) i.e. by white line closure, or any other programmatic closure?
2/ any idea why we might not see closes in an MT4 backtest?

I'm assuming it's a bug, although I haven't been able to spot it yet. DoesTradeExist() should select the right trade, CloseTrade() should then close it when it's called (straight after).

latest version attached - for debugging. This one does contain one logic program with the strategy - reversal trades can be left without exit critera under a couple of circumstances. I'm thinking through what the best options are here and will code up. In the meanwhile if anyone can spot the Trade closure problem I'd be grateful.

cheers
Simon

if i change the EmergencyStopLoss to 9999 then it doesn't do the Buy trades for U/J with AVAfx
When I change it to 9999 With IBFX I get;
Nanningbob V5 auto-trader by Steve and Simon USDJPYm,M5: USDJPYm 0 SL order modify failed with error(4051): invalid function parameter value

.. but do get both Buys and Sells, EmergencyStopLos probably set way to high

Martingaling, the key is how you go about doing it, and deciding when you've gone too far and its time to give up.

So you have the traditional martgingale:

1, 2, 4, 8, 16, 32, etc..

Which makes sure you always end the last trade in profit.

The Nanningbob 242 System:

1, 2, 4, 2,

Which I'm not entirely sure about, but isn't 100% to have the last trade in profit or break even.

And the Sum Form:

1, 2, 3, 6, 12, 24, etc..

Which makes sure you always end the last trade at break even. The Sum Form version being the least risky in my opinion.

Choosing the style to use depends entirely on the system it will be applied to and how often it experiences a string of losses. Its also important to keep in mind the scale of the losses vs the wins. For instance if your wins are generally 2x larger then your losses, then perhaps you can get away with a fractional martingal to cover and eliminate the losses.

For instance that EA I posted earlier based on the WPR 6th, if you backtested it on a 4hr chart like I did and saw that it had 98% loosing trades, yet was still profitable.. would it be a good martingale candidate? Hell no! Because when you loose 50 trades in a row your lot size would be so ridiculous any hope of profit would be forever lost. However what made it profitable was that when it did have a win it was a large win, and the losses were very small.

Good EA's for a martingale system are EA's that already have a 50% or higher win percent. So if you backtest your EA with no martingale in place and see you win on average 51% of your trades and your losses and wins on average are the same size, then implementing martingale may be a good idea.

Perhaps I'm singing to the choir here, or maybe I've made you think about it in a different perspective.

I maintain there is nothing inherently "bad" about martingale; its just been implemented poorly in a lot of EA's and trading systems to make up for the trading systems inherently flawed entry and exit signals, which ultimately leads to a string of losses and account death. Once it got a bad reputation through poor use, its hard to get people to change their minds and look at it in a positive light again.

Have you seen this contribution from arbinka - http://www.forexfactory.com/showpost...postcount=992?

Arrg. I'm an idiot. That's it. Thank you Arbinka. Thanks Steve.

I'll fix tonight, test and post.

cheers
Simon

Hehe. I am sooooooooooo familiar with this feeling.

There is a microbe in the screen display - it does not show the updated magic number when in 2.4.2 recovery mode. This is only a display thingy, so download the fix only if you feel like it.

Coders, find this line:
and replace it with


Non-coders can do this as well, if you have read up on changing script codes - just remember to re-compile when you have made the change.

I like being humbled. It usually means I'm making progress!

On the progress front, I have now eliminated most of the runaway loser trades... sadly only most. there are a couple of other conditions I need to figure out. Won't be posting it up tonight.

The major issue are what are the best exit conditions for failed with trend, and failed *reversal* trades. NB uses 2.4.2 recovery for those recovery trades that go wrong... I'm flying out into unknown territory in figuring out when to decide a reversal isn't going to work and when is the best time to close it.

cheers
Simon

How abut using a percentage of the account equity to determine the reversal point ... so 1.5% for the reversal to kick in, an 2.25% for the close. (figures are arbitrary and could be user selectable)

This is where is starts to get complicated. I am guessing it should close one and open the other.

Is there a possible glitch within the original nanningbob code? Why was the T/P set lower than the price it entered? I haven't touched this bot all week and just letting it run its course. Oh and I'm using bollingerband T/P with all other T/P off. Same settings that I posted a few pages back too. This is the current version of nanningbob v5 on page 1.

http://i148.photobucket.com/albums/s...g?t=1282685329

It looks like your Recovery method Kicked in .... 0.2 lots on the second trade brought you to breakeven and the first trade exited out at the same time sto zero you out

Yep. That is 2.4.2 kicking in and hauling your butt out of the fire.

Yep. That's a bug. Shouldn't happen. I've been using the backtest to iron out the obvious errors across a wider range of scenarios. At this stage i'm still not sure if my variant is A/ profitable, B? better than straight NB v5.3. Chances are it won't be better, because Bob has developed a great system, and had years to refine it.

But.... even if what I'm doing fails, I think because of this experiment, I'll be able to add a minor tweak to NB v5 which should improve its profitability and reduce its drawdown. That was the original aim, and I think still achievable.

Right now I'm working on restricting with trend entries to only those that are "BB huggies" by tightening MinSepfromBB, and I'm about to change the with trend exit to be when price separates from the bollinger band by a certain percentage. This usually is a precursor to a reversal. The other condition will be NOT to enter a reversal trade if in a with trend trade.

Still a way to go I'm afraid.

cheers
Simon

It's looking good, you are right, regardless the end result could be a reduction in the DD in NBv5. Your version i gotta say, is a monster when it comes to opening trades. This bad boy will trade just about anything!!!

I running NBTS v1.00 and just checking in for the week to see ho it's going. I find many many error messages: "order send failed with error (130): invalid stops".

I guess I should just change to v1.01...or an I missing something in the setup of the EA.

Cheers
CanuckCT