stevehopwoodforex.com • Zigzag Bollinger Band

Author:

snailbeard [ Wed Jun 11, 2014 2:05 pm ]

Post subject:

Adaptive stop-losses

In the previous post the swing-following jumping-stop method successfully grabbed lots of pips but the market always kicks back with some nuance to undermine a fixed approach. In this chart, we see how difficult it is to use non-adaptive methods. After being in profit by about 150 pips, most of it slips away leaving just about 20 pips!

Consider alternative methods of using stop-losses:

1. Go back to a smaller fixed target (wasteful in strong swings)

2. Calculate the target based on weekly pivots: e.g. (WR1: 80 pips) and the expected range D1-ATR is 90 therefore 80 pips is cautious but also dynamic, WMR2 is more ambitious at 120 pips, some times achievable in a single day, in this case it took two days, sometimes even WR2 is reached but more information and analysis would be required.

3. A fixed proportion stop-loss is another method e.g lock 50% of profit after n-pips in profit, which is also a wasteful approach but better than 1.

4. I notice that the H4 jumping-stop-loss is often quite good if it isn't used too soon, but would be even better if H4-ORFB entry/re-entry methods were reliable enough to get back into the swing after profit taking.

In summary, I like the idea of a daily ATR target which is linked to the nearest weekly pivot level.

In this example:
The step of WMR1 to WR1 is 66 pips, if locking 20 pips: 20/66 = 30% allowance and this is signifcantly better than previous experiments with 50% locking which is often very wasteful.

In the second stage trade:
WR1 to peak price: 73 pips, and if using 30% locking then 22 pips of profit leakage allowed to the SL gives a similar performance.

This suggests that when price is close enough to a profit target
we can use 30% of profit pips as the trailing stop loss method,
and a take profit exit if price reaches the appropriate pivot.
We might also prefer to use a SL which is less than the TP so that the TP(pips)/SL(pips) > 1

To switch SL mode during a trade on a specific trade using MetaCrap MQL can be a bit messy, but it works for the switch from D1 SL to H4 SL, so it doesn't require too much new code.

old result: 67 2012.01.19 13:10 buy 15 0.30 1.04172 0.00000 0.00000 0.00 9061.96 68 2012.01.19 13:10 modify 15 0.30 1.04172 1.03272 1.07864 0.00 9061.96 69 2012.01.22 22:00 modify 15 0.30 1.04172 1.03830 1.07864 0.00 9061.96 70 2012.01.24 00:00 modify 15 0.30 1.04172 1.04372 1.07864 0.00 9061.96 71 2012.01.24 13:41 s/l 15 0.30 1.04372 1.04372 1.07864 61.80 9123.76 new result: 1 2012.01.19 13:10 buy 1 0.30 1.04172 0.00000 0.00000 0.00 10000.00 2 2012.01.19 13:10 modify 1 0.30 1.04172 1.03272 1.05064 0.00 10000.00 3 2012.01.22 22:00 modify 1 0.30 1.04172 1.04532 1.05064 0.00 10000.00 4 2012.01.23 08:18 t/p 1 0.30 1.05064 1.04532 1.05064 268.80 10268.80

05:15 Alert: bZzM15PentaH4Triplet && (iZzTH4PM15Dir > 0) 05:15 Alert: ZZGoodForBuy bCloseAtM15MBB && bM15H4TriggerConfirmed 05:15 Trade because: (bHasIntraDaySlope && bBreakH4ORFB) : 1 05:15 Entry Dropped: FAILED (W1) checkW1FiboH4CloseOK

Author:	snailbeard [ Sun May 25, 2014 6:32 am ]
Post subject:	M1 Price Action versus ADX
The M1 view shows the market action unfolding and we really see the pips slipping away on this chart. Price action can be defined generically as (pips moved / time passed) although it could defined as price passing several price levels or a large candle body after smaller candle bodies. In this paticular case, we get a sequence of M1 bars breaking out of the previous range in our prefered direction with increasing momentum:

Author:	snailbeard [ Wed May 28, 2014 6:19 am ]
Post subject:	The trend is your friend...
Only it can be difficult to see it before it is nearly over. I prefer to say swing is king, because the trend is important on a higher time frame but we need to spot an emerging swing back into the trend before that swing has gone too far. I recently reworked trend-swing analysis to make use of multiple moving averages which I have done in the past without consistent results. The lagging nature of even fast MAs means that there will often be times when the market will swing at just the right frequency to trigger trades so that we get a sequence of losing trades. To avoid the lagging issue I have implemented a stack of MAs fast and slow on several time frames and this often produces good interpretation of the trend and swing, however, some of the best entries occur as price moves back to the level of the slower MAs and then rejoins the longer term trend. unfortunately even the fast MA's cannot change fast enough to produce the correct directional bias in time. I am now compenstaing for this by scoring 'flatness'. This score gets higher as more MA's come together and peaks one H4 bar before the fast MA's get back into the swing direction. Since flatness will also occur at a reversal I look at the candle patterns to find out if price action is going with the main trend. In the attached chart we can see a new swing emerging after a pullback and H4 candle bias confirms the dominant direction:

Author:	snailbeard [ Thu May 29, 2014 6:32 am ]
Post subject:	ZzBb matures
It has been a long hard slog to get here.. It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of Light, it was the season of Darkness, it was the spring of hope, it was the winter of despair... Despite the difficulties of working with MetaCrap and CrappyTester, perseverance might be finally paying off. ZzBb survived three years of historical data and made a profit. However, to make the best profits it needs to be switched on when D1 is trending and switched off when it isn't. In the first chart we see how it performs for a trending pair over two months:

Author:	snailbeard [ Thu May 29, 2014 6:37 am ]
Post subject:	ZzBB should not be run during D1 ranging periods
If ZzBb is enabled when D1 is ranging it will result in lots of losing trades:

Author:	snailbeard [ Fri Jun 06, 2014 9:07 am ]
Post subject:	Experiments with jumping stoplosses
Experienced traders advise us to let our profits run. The previous version of the EA was running with a fixed target and fixed stoploss until recently. The result was a lot of missed pips and some winning trades going back to the stoploss. In the current version and this example, the main swing lasts for about a week before there is a significant pullback. Although, you can easily identify two obvious entry and take profits instead of a single trade. If I could reliably get the timing right for the second entry then I suspect the long term results would be better. This example, which is almost ideal, might not be typical. The entry occurs early due to rapid price action at 18:30 on the 5th. However, a more reliable entry does not occur until the next day. There have been 3 breaks of H4 mid-Bollinger band. The H4 bar which opens at 4am on the 6th with a ORFB (Open-Reverse-Forward-Break) is perhaps the most convincing break of the MBB threslhold. Splitting the trade into two trades would require setting the first target to WR1 and the second target to WR2. The second buy entry has to occur before EMA5 (red) resumes a positive slope, therefore we cannot use the same entry logic without being late into the second part of the trade. So in this particular case it makes sense to let the profits run. I have added two varients of a jumping stop method. At the beginning of the trade the stoploss follows the low of the previous D1 candle, then there is a change to following the low of the H4 candle when price moves closer to the target price. It works really well when there is enough momentum to reach WR2 but would be very wasteful if the target was too ambitious. Over the long term it might be better to take profits at WR1. One can also combine jumping stoplosses with active exits after reversal in some of the EMAs.

stevehopwoodforex.com https://www.stevehopwoodforex.com/phpBB3/ Print view

Zigzag Bollinger Band https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=3237	Page 18 of 28

Author:	snailbeard [ Mon Jun 16, 2014 4:12 pm ]
Post subject:	Better profit locking
I have modified the take profit calculations to be more focused on the daily ATR and enhanced the trailing stop-losses. There are now three modes of profit locking via moving stop-losses a) Initially a large stop-loss based on the extreme of the previous day b) On approaching the target price (say two thirds of the target) switching to H4 jumping candle stop-losses c) On getting quite close to the target switching again to percent profit method of trailing, thereby locking about 75% of the profit. However, the H4 candle stage might be bypassed when the price moves fast enough towards the target, which happened in this case. We can see the difference this makes to the result of the previous post: Code: Select all old result: 67 2012.01.19 13:10 buy 15 0.30 1.04172 0.00000 0.00000 0.00 9061.96 68 2012.01.19 13:10 modify 15 0.30 1.04172 1.03272 1.07864 0.00 9061.96 69 2012.01.22 22:00 modify 15 0.30 1.04172 1.03830 1.07864 0.00 9061.96 70 2012.01.24 00:00 modify 15 0.30 1.04172 1.04372 1.07864 0.00 9061.96 71 2012.01.24 13:41 s/l 15 0.30 1.04372 1.04372 1.07864 61.80 9123.76 new result: 1 2012.01.19 13:10 buy 1 0.30 1.04172 0.00000 0.00000 0.00 10000.00 2 2012.01.19 13:10 modify 1 0.30 1.04172 1.03272 1.05064 0.00 10000.00 3 2012.01.22 22:00 modify 1 0.30 1.04172 1.04532 1.05064 0.00 10000.00 4 2012.01.23 08:18 t/p 1 0.30 1.05064 1.04532 1.05064 268.80 10268.80 Of course the trade had further to run and it would be good to piggy back a couple more trades on this one by ORFB on H4, but I doubt that will happen/work using MetaCrap MQL.

Author:	snailbeard [ Tue Jun 17, 2014 4:17 pm ]
Post subject:	Two hits and one miss
After improving the Fibonacci entry filter (which was blocking T1), ZzBb reclaims the missing trade, so both T1 and T2 go well, in addition, there is a potential trade at T3, but the fast-EMAs are too lagging to get the swing in time for another trade. ZzBb needs a Price-Action override so that we can ignore the direction of the EMAs and trigger by ORFB instead. It appears that the main T3 price surge is over in a single H4 bar, and the crossing-with-strength of WR1 could be an appropriate trigger.

Author:	snailbeard [ Wed Jun 18, 2014 7:33 am ]
Post subject:	More Fibonacci issues
After investigating the missing T3 trade and following the log file bread crumbs, it appears that ZzBb found a precursor entry (05:15) before the main price action (after 16:00), started setting up a trade only to have it thrown out by the Fibonacci filter: A potential trade is picked up by both the Zigzag and ORFB entry methods: Code: Select all `05:15 Alert: bZzM15PentaH4Triplet && (iZzTH4PM15Dir > 0) 05:15 Alert: ZZGoodForBuy bCloseAtM15MBB && bM15H4TriggerConfirmed 05:15 Trade because: (bHasIntraDaySlope && bBreakH4ORFB) : 1 05:15 Entry Dropped: FAILED (W1) checkW1FiboH4CloseOK` It looks like there is another defect in the Fibonacci filter, perhaps I'm using the retracement levels instead of the expansion levels. In the H4 chart, the thick blue lines correspond to Fib. 100 and Fib. 161.

Author:	rasta420 [ Thu Jun 19, 2014 11:06 am ]
Post subject:	Zigzag Bollinger Band
excuse me sir,Where are your EA

All times are UTC	Page 18 of 28
Powered by phpBB® Forum Software © phpBB Limited