Slopey Peaky Bob

[thomas1386]

There is a new version of the SPB Core Library.

Alex (biobier) emailed yesterday with a bug spot that stops the trailing and jumping stop functions from working except on certain pairs.

This has not come up as a problem before I think, because we have been basket trading. It is only of relevance to you if you are doing any single trading and want these two functions.

What you need to do to access the fix is:

• Download the library into your platform's /Include folder.
• Go to your platform's /MQ4/Experts folder and delete the mqlcache.dat file - it has a small back bix with a white dot before its name.
• Shut down and restart the platform.

It is an easy DIY:
- Go to: void jumpingStopLoss(int ticket).
- Scroll down a few lines to double sl = OrderStopLoss();
- Insert this code underneath:

Code: Select all

//This next line fixes a bug spotted by biobier. Thanks Alex.
  double point = SymbolInfoDouble(OrderSymbol(),SYMBOL_POINT);

- Go to: void trailingStopLoss(int ticket) (next function down)
- Repeat the insertion.
- Go back to the top of: void jumpingStopLoss(int ticket)
- Do a search and replace, replacing "Point" with "point". There are two instances in each function.

This bloop also applies to Trendy Monster Moving Day and Desky, so I will post updates there.

Thanks Alex.

Hello Steve,

I have noticed when setting the basket take profits in pips it is not shown correctly by SPB. The individual pairs takeprofit only shows 0 Euro and no pips. Whole basket is shown correctly. When I set both to cash it also shows both correctly.

Greets, Thomas

[dclayw]

My current 5 SPBv2s tests have been running for 10 weeks. All these tests are on GP and as I'm in Australia I have them running at 1:30 leverage which is maximum allowed now, all $20k a/cs trading 0.01 lots. To limit the number of trades and avoid margin call I'm using forexkiwi set to 450 so no more trades are created once margin % drops below that. With this setup the test gets to around 100 trades or so then stops initiating any new trades. (I changed forexkiwi to 500 2-3 weeks back). I also have Shirt Protection settings at 300,125.

The 2 best performing tests do not use peaky at all, just different setting of SS. The best performing one was up 15% in 10 weeks and looking good. I've attached the set file. It uses a grid of 4 but also sends immediate market orders as this strategy won't work without that. This test reached 100 trades or so many weeks ago and has just been cruising along, not initiating any more trades but slowly grinding further into profit on most pairs. The pendings were also slowly being activated and whittled down. Margin % down around 450 is getting into tricky territory but the protection settings seem to be working quite well. With my SS setting on this test the SS does not change colour often (spends a lot of time in neutral territory).

Today when I checked it I saw massive numbers of the "order send failed with error (2): common error" and trades on about 8 pairs were continually being sent and returning this error. I am not sure why the EA started sending trades again. Forexkiwi is at 500 and I'm certain that level would not have been reached during the day. Margin % has been hovering around 430-450 for weeks. All the profit was on open trades and GP eventually shut the account down and killed them all, deleting most of Account History. Also getting the "AccounHistory cannot login to GlobalPrime-Demo" now but that's just because GP shut down the account I guess.

A little bit disheartening having a test performing so well and then the EA going nuts. Something changed but not sure what. Not sure where to look. Anyone experience similar with SPB2s? I ran a series of SPB2s tests for 6+ months a while back with not a single issue.

[trader689]

hey dclayw

I had the same thing happen (same error message) to me across a number of my Global Prime demos yesterday, and not just restricted to EAs from this forum. A couple were from completely different EAs, and a couple of Tdesk demos I have runnng (all of which, similarly frustratingly, were doing very well)

I got in touch with Global Prime's (excellent) support, and they said it was because of hyperactivity that they shut the account. Any open demos you have which are currently showing the error (2) message I'm afraid may well be 'zombie' demos. What I mean by that is they look like they're alive and connected to the server from the icon on the bottom right, but if you save the account number and password from the mailbox tab, and try to login again, they may show as 'account invalid' - i.e. a deleted demo.

I suspect it might have been because of the huge downspike at market open yesterday on GBPAUD and GBPNZD (and perhaps other) pairs, as these events dovetailed, but can't be 100% sure

trader

...Today when I checked it I saw massive numbers of the "order send failed with error (2): common error" and trades on about 8 pairs were continually being sent and returning this error. I am not sure why the EA started sending trades again. Forexkiwi is at 500 and I'm certain that level would not have been reached during the day. Margin % has been hovering around 430-450 for weeks. All the profit was on open trades and GP eventually shut the account down and killed them all, deleting most of Account History. Also getting the "AccounHistory cannot login to GlobalPrime-Demo" now but that's just because GP shut down the account I guess.

A little bit disheartening having a test performing so well and then the EA going nuts. Something changed but not sure what. Not sure where to look. Anyone experience similar with SPB2s? I ran a series of SPB2s tests for 6+ months a while back with not a single issue.

[dclayw]

hey trader689 thanks for the feedback, good to know it wasn't just me!, and not restricted to just SPB either. I did contact GP as soon as I saw the problem and yes they said the same thing, they shut it down due to hyperactivity. I am running 10 demos atm across 2x VPS, 5x SPBv2s and 5x WPB and only this one had the problem (the best performing one ofcourse). The other 4x SPB tests still cruising along.

I see the spikes you're talking about, it's a bit of a worry if something like that caused the issue. My test that had the problem should not have been initiating any new trades at all (and had not initiated any for a few weeks), because margin % had dropped below forexkiwi setting. It's hard to know what caused the problem, whether it's the EA itself under certain conditions or the broker platform or what. I used to use the MinMinuteBetweenTrades setting, might have to go back to that.

SPB was sending the same trades over and over for about 8 pairs, every 10 secs. I am not sure why every 10 secs because I have PostTradeAttemptWait set to 60s secs, so the EA should have waited at least this long before attempting to resend. I do have the EventTimerInterval set to 10 secs so it looks like trades were being resent on that interval.

Did this happen to anyone with a live account?

[trader689]

hey dclayw,

my instinct was also that although a sharp spike, it didn't seem enough of a turmoil to precitipate hyperactivity from the EAs on some of the demos which ended up being shut down. Yes it is a tad worrying, I don't know whether something technical needs examining at Global Prime's end. Hopefully we won't see similar going forward, but let's continue to share experiences

trader

[tomele]

I second that.

Out of 29 running demos, 4 got shot down on my side. These were totally different EAs and I couldn't find anything unusual in their logs. Other demos with identical EAs were unaffected and in part had much bigger logs, i.e. they were much more active and are still alive.

It is a bummer as some of these demos were running for years without any issues. The randomity of the affected tests speaks volumes. There can always be technical issues like defective drives that everybody would understand. Blaming hyperactivity seems not comprehensible in this case.

On the other hand, we owe GP so much for being the honest broker they are and for offering the capacities to demo test all our weird ideas without time limits. Let' swallow the anger and move along, please.

[trader689]

hey tomele

many thanks for sharing your experiences as well, and it seems to dovetail with mine and dclayw's observations too. Totally agree that we have so much to be thankful for to GP.

The aim of the discussion, at least on my part, is not about being angry with Global Prime. It is to potentially highlight a possible error at their end, and if feedback is given to them, especially by such a credible member such as yourself, I'm sure that they will appreciate it as it will help to highlight a possible vulnerability / loophole on their end. It will also reassure GP users (I think the vast majority of us here) that going forward, demo accounts, some of them long-running, from which traders are drawing vital decisions on what systems they would consider trading live, will not be randomly deleted

Is there a case to reach out to Jemook and broach this issue with him?

trader

[taipan]

I second that.

Out of 29 running demos, 4 got shot down on my side. These were totally different EAs and I couldn't find anything unusual in their logs. Other demos with identical EAs were unaffected and in part had much bigger logs, i.e. they were much more active and are still alive.

It is a bummer as some of these demos were running for years without any issues. The randomity of the affected tests speaks volumes. There can always be technical issues like defective drives that everybody would understand. Blaming hyperactivity seems not comprehensible in this case.

On the other hand, we owe GP so much for being the honest broker they are and for offering the capacities to demo test all our weird ideas without time limits. Let' swallow the anger and move along, please.

I also second that.

3 of demo aslo being halted by GP due to hyper activities.. Restarted to register the new demo in GP clinet portal but could not be connected to GP demo server. It could only be connected after sometimes.

taipan

[SteveHopwood]

Thomas and I reached the same conclusions pretty much simultaneously.

Do a quick detour to this post to read about the doubling of the account balance: viewtopic.php?p=170033#p170033

I had a second demo running; that also doubled in size, so I had two accounts doubling from $10,000 to 20,000 plus.

Both demos gradually accrued huge positions over the succeeding 12 months, positions that did nothing more than cost vast amounts of swap costs. I finally got fed up a few weeks ago and closed them both.

This still left an excellent profit with balances at $11,750 and $11,634 respectively. We would be hard put to make these returns leaving the money in a savings account.

On the other hand, the positions would have blown a $10,000 account had I started the demos on April 20th 2020.

Our conclusions are:

• Timing is everything. Start taking trades at the right time and everything is fine and dandy. Hit a 'leaky toilet' market though, and SPB gets killed.
• This is not a safe EA.

For those of you unfamiliar with the term, a 'leaky toilet' market is nanningbob's description of a market that moves slowly but inexorable against our trades, i.e. falls against buys. The retraces are brief and not large enough to allow orders to be closed without loss.

The two demos are still running and one has already closed a basket. I shall leave them running.

[dclayw]

Yes this past year has been somewhat disappointing as SPB has been the most promising EA for some time. Then again not too many strategies work well in such low volatility sideways markets as we are seeing. I guess the problem is when to use it and when to leave it alone. With long term strategies it's a difficult question to answer.