• CloseTradesOnRelevantSemafor: closes buy trades if the previous candle closed with a high semafor. Closes sell trades if the previous candle closed with a low semafor.
• CloseImmediatelySemaforAppears: as above but closes immediately the signal appears. You can have both CloseTradesOnRelevantSemafor and CloseImmediatelySemaforAppears.
• CloseOnlyInPositivePips: this tells HGBnG only to close a group of trades if they are in pips profit as a group.
• DeleteAndReplaceLosersGrid: Losing hedged trades will be a long way away, so delete outstanding stop orders and replace them closer to the market.
• The remaining inputs tell HGBnG how to behave with regard to replacing a grid after a semafor-signalled group of trades are closed. For the purpose of this description, imagine a high semafor has closed a group of buys (mentally turn everything upside down for sell closure):
  • ReplaceWinnersGrid: tells HGBnG to send a new group of buy stops to catch the trend resumption.
  • ReplaceWinnersByLimitGrid: if the closed buys were a part of a hedged position, this tells HGBnG to send a group of buy limit trades at DistanceBetweenTradesPips below the market. This was Kevin's idea for dealing with the situation where the market retraces briefly after the new high is reached, then resumes its upward climb. It cuts the DD created by the existing market sells.
  • AddStopOrdersToLimitGrid: there will come a point where the uptrend will end. The market will fall, filling all those limit orders leaving market buys in a selling market. This tells HGBnG to adopt Pat's idea of adding a grid of sell stops below the buy limit grid. This will add more sell trades to the basket as the markets fall, eventually allowing the position to close at breakeven and banking all the pips made whilst the market rose.

k3v1n0s » Mon Mar 14, 2016 12:20 am wrote:

patmontes » 13 Mar 2016, 14:50 wrote:Just thinking out loud... (and willing to be shot down )
Is it worth sending at the same time already the sell stop orders, X pips away below the buy limit orders so that when it goes wrong, we already are on our way closer to the target Global profit given that the trend is reversing and the semafor high did not repaint?

Excellent idea Pat! This could make HGB'nG a hands-free EA again.

Seems like Steve is already making the changes to the code.

Can't wait for tomorrow.

Thank you very much Pat!

Cheers,

Kevin

SteveHopwood » March 12th, 2016, 10:15 am wrote:Courtesy of Kevin, V 2j is in post 1. What astonished me is that he appears not to need a bath before doing useful thinking.

Hang on to your socks folks, 'cos this is going to blow them off. I am going to let him explain in his own words, so here is the pm he sent me. Take your time reading it:

-------------------------------------------
"Hi Steve,

What amazing results we had with HGB'nG this week.

There's no doubt that HGB'nG has the potential to be one of the most profitable EA's.

Still, after this week's events, I noticed a dangerous downside to the way we are running HGB'nG at the moment. I'll explain why, and trust me, I got a potential solution to it (already coded).

After Thursday's events, I was very euphoric as I saw that my balance was growing rapidly and my equity was following closely with low drawdown (< 10%). But on Friday, as my balance kept growing, my equity fell to a drawdown of 25%, eating on my all time HGB'nG's profits. I investigated why this was happening, and found out that most of that loss is caused by my 10 open GJ sell trades. As high semafor signals were appearing, my buy trades were being closed out, new sell stop trades were placed, some of them filled, then price headed north again, this cycle repeated, creating a ever growing drawdown.

This is not dramatic, as we are intelligent beings and so I started to think of a way to get out of that big drawdown. What I noticed is, that although semafor signals are very good areas to close our trades out, they almost never (in real-time, not in hindsight) act as good reversal signals. You already know it, but plotting the Quantum on a chart shows us that these signals keep appearing in a downtrend as well as in a uptrend :





So my rhetorical question is : Why the heck do we want to be sellers in an uptrend and buyers in a downtrend? If we know that high semafor signals are more an indication of an uptrend than a downtrend, then placing sell orders on that signals, is just going to act as a silent (but fast) death on our equity. If this GJ uptrend continues with the same momentum as has been since wednesday, then if nothing changes in the way I am trading HGB'nG, my real account will be wiped out by end of next week.

Now comes my suggestion. Why don't we add an option, that when a high semafor signal appears, makes HGB'nG simply close out our buy trades that are in profit and sends a new grid of buy stop orders and a new grid of buy limit orders? This would allow us to catch the retrace if there is one, profit from the uptrend, and if the retrace is deep enough and we catch enough buy limit orders, it possibly would close out our whole position when "HedgeProfitPips/Cash" is attained, ending the cycle.

If it still goes wrong, and the retrace is in reality a reversal (so the high semafor was really a 3 level high semafor which won't repaint), then we end up with a position that has to be closed out manually, but hopefully with a global profit. Looking back on many charts, real semafor signals are very rare compared to quantum signals. I strongly believe that this strategy would work better then adding to our losing position and accumulating drawdown, but whoever wants can turn the limit order option off. We can also later as we see how the EA performs on this, work out a good way to handle those buy orders that happen to open on the last semafor signal.

Steve, what do you think of this? As this text is getting pretty long, I'll write another PM with the changes I've made on the code.

Cheers,

Kevin"
-------------------------------------------

Stunning, or what? You will find the relevant inputs in the "3 semafor inputs" section:

• ReplaceWinnersByLimitGrid tells HGB to behave as described.
• DeleteAndReplaceLosersGrid tells HGB to behave as it did previously.

Default inputs are effectively my set file.

This does move us a step away from having the holy grail of an ea that can be left alone, but anyone here seriously thinks that Kevin will not come up with ideas to improve the bot still more once the markets reopen and we start trading again.

Fantastic stuff Kevin. Absolutely brain-blowing.

I bloody hate weekends.

SteveHopwood » Mon Mar 14, 2016 9:38 am wrote:V 2l is in post 1, with two teensy but important changes that you can make for yourselves if you want.

Line 3919
SendSellGrid(Symbol(), OP_SELLSTOP, NormalizeDouble(Bid - ((DistanceBetweenTrades / factor) * GridSize), Digits), Lot);
becomes
SendSellGrid(Symbol(), OP_SELLSTOP, NormalizeDouble(Bid - ((DistanceBetweenTrades / factor) * (GridSize + 1)), Digits), Lot);

Line 3968
SendBuyGrid(Symbol(), OP_BUYSTOP, NormalizeDouble(Ask + ((DistanceBetweenTrades / factor) * GridSize), Digits), Lot);
becomes
SendBuyGrid(Symbol(), OP_BUYSTOP, NormalizeDouble(Ask + ((DistanceBetweenTrades / factor) * (GridSize + 1)), Digits), Lot);

Then line 8 to #define version "Version 2l" if you want the chart displaying the correct version no.