Pair Divergence Method | Page 31 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

What To Pair Divergence With? 14 replies

Divergence and hidden Divergence 122 replies

bo7a method... method for GBP/JPY 205 replies

Divergence method 11 replies

journal for pair divergence method 33 replies

Attachments: Pair Divergence Method
Exit Attachments

Pair Divergence Method

  • #601
  • Oct 4, 2009 10:37am Oct 4, 2009 10:37am
My position on the 'wiggles': if the wiggles are correlated, go for it.

What is the market trying to tell us? If you see something that makes you think twice, but you still think you know, then go ahead and pull the trigger. You will find out if you were right or wrong in judging what the market was saying, and the sum of these experiences will help you to become a better chartist.

Enough zen talk for now

Quoting stevegee58
If you think about it, if you have a divergence with wiggles in between on one time frame, it'll be a divergence with no wiggles on a higher time frame. Get in the habit of scanning the M15 through H4 time frames regularly to get a feel for it.
I agree.
Hindsight is 20/20
  • #602
  • Oct 4, 2009 12:42pm Oct 4, 2009 12:42pm
The following link might be of some value to you: Click here
Quoting stevegee58
Hey Bill, I can't think of a reason why this wouldn't work with other correlated markets. For instance ES/YM (index futures), or maybe 2 liquid stocks in the same sector like Exxon and BP.

I put ES and YM on the same chart on Friday to watch them but they're so perfectly correlated that there weren't any divergences.
  • #603
  • Edited 12:54am Oct 5, 2009 12:09am | Edited 12:54am
I just finished with a visual back test of August 09, where I only traded the hours of midnight-noon, EST. I didn't trade any 'flat' divergences. I scanned through the month on the 15 minute time frame, glancing at the 1 hour every now and then. I then scanned through the month on the 1 hour time frame. I don't really feel I missed anything by not scanning the 30m -- who knows. I allowed all valid entries, even if it doubled my risk at the time. In the event of a close call on SL/TP, I allowed room for the spread, but I think this was only an issue on two trades.

14 wins
7 losses
1 trade reversed for break even (trade triggered, entered, another triggered for opposite direction on the same pair. If you would have ignored the second one, the first would have been a winner)

If you would have traded at 3% account equity per trade, your account would have grown by 22.2% (less spread and swap)
If you would have traded at 2% account equity per trade, your account would have grown by 14.5% (less spread and swap)
If you would have traded fixed lots, you would have gained 178 pips (less spread)
Hindsight is 20/20
  • #604
  • Edited 11:16am Oct 5, 2009 10:31am | Edited 11:16am
U/C buy on the 15m with ATRx2 = 22 pips.

edit: this trade stopped out for a loss.
Attached Image (click to enlarge)
Click to Enlarge Name: eurusd_buy_200910051700.gif Size: 16 KB
Hindsight is 20/20
  • #605
  • Oct 5, 2009 2:37pm Oct 5, 2009 2:37pm
Quoting cosgrove
U/C buy on the 15m with ATRx2 = 22 pips.

edit: this trade stopped out for a loss.
i took this trade too
  • #606
  • Oct 6, 2009 8:16am Oct 6, 2009 8:16am
I can't edit my old post, but the U/C sell on the 1h marked on the October 2nd chart as 20091002 @ 15:00 resulted in a winner.
Hindsight is 20/20
  • #607
  • Edited Oct 7, 2009 8:35am Oct 6, 2009 8:55am | Edited Oct 7, 2009 8:35am
Euro buy on the 30m. Almost stopped out but still hanging in there.

edit: after almost being stopped out again, it hit profit.
Attached Image (click to enlarge)
Click to Enlarge Name: eurusd_buy_200910061030.gif Size: 18 KB
Hindsight is 20/20
  • #608
  • Oct 7, 2009 10:47pm Oct 7, 2009 10:47pm
there is a shot on EURUSD right now. 15 min

EDIT. billbs, you will never look at S&R with this method? Because with this setup for example, we are fading a breakout with good momentum.

another question, you are not trading 5 min anymore?

thanks in advance.
  • #609
  • Oct 7, 2009 10:49pm Oct 7, 2009 10:49pm
i hope lucid came back with a new indicator. The biggest barrier for test this method now is the indicator. Changing scale untill se the two line together make the lines almost flat, it is a pain.
  • #610
  • Oct 8, 2009 8:11am Oct 8, 2009 8:11am
Quoting giraia_br
there is a shot on EURUSD right now. 15 min
For the record, I'm only trading between GMT 0100-1300, as this would be the time frame in the future i'm willing to be available for trading

Attached a winning U/C buy earlier this morning. 5 decimal places makes it a tough call as to when you would consider it higher low/lower high. I rounded the values of the E/U MA on an E/U chart, and there was a one pip difference so I counted it (the U/C higher low was obvious, the E/U lower low was arguable).
Attached Image (click to enlarge)
Click to Enlarge Name: usdchf_buy_200910081000.gif Size: 17 KB
Hindsight is 20/20
  • #611
  • Oct 8, 2009 2:56pm Oct 8, 2009 2:56pm
U/C 1hr buy for a loss. Entry bar was huge, not sure what you'd want to do with it.
Attached Image (click to enlarge)
Click to Enlarge Name: usdchf_buy_200910081500.gif Size: 18 KB
Hindsight is 20/20
  • #612
  • Oct 8, 2009 3:05pm Oct 8, 2009 3:05pm
Quoting giraia_br
there is a shot on EURUSD right now. 15 min

EDIT. billbs, you will never look at S&R with this method? Because with this setup for example, we are fading a breakout with good momentum.

another question, you are not trading 5 min anymore?

thanks in advance.
I never look at SR. With this method, if we worried about SR, pivots, % of daily range etc.. we would wind up with one trade a month and it would probably lose.
  • #613
  • Oct 8, 2009 3:08pm Oct 8, 2009 3:08pm
Quoting cosgrove
U/C 1hr buy for a loss. Entry bar was huge, not sure what you'd want to do with it.
I wouldn't have taken that trade even with a normal bar.
Look at the aqua pair. You can't draw a trendline connecting the two vallies without hitting the middle wave.
  • #614
  • Oct 8, 2009 3:13pm Oct 8, 2009 3:13pm
Quick visual check of early September 09 combined with the results I had posted so far in the thread led me to these stats (could be off by one or two, but I hope not)

9 wins
4 losses
2 trades combined for BE (one entered, got a signal for the opposite direction, so I got out of the first at BE or close enough)
Hindsight is 20/20
  • #615
  • Oct 8, 2009 3:15pm Oct 8, 2009 3:15pm
Quoting billbss
You can't draw a trendline connecting the two vallies without hitting the middle wave.
I agree, and I went back and forth with myself over whether or not it should be a valid trade. I probably won't trade these in the future.
Hindsight is 20/20
  • #616
  • Oct 9, 2009 11:20am Oct 9, 2009 11:20am
Euro buy this morning; hit profit.
Attached Image (click to enlarge)
Click to Enlarge Name: eurusd_buy_200910090915.gif Size: 15 KB
Hindsight is 20/20
  • #617
  • Edited 8:39pm Oct 9, 2009 4:07pm | Edited 8:39pm
Just did a visual backtest of July (1h, 30m, and 15m timeframes, only taking trades between 06:00-18:00 GMT -- totally misspoke earlier in the thread and said 0100-1300 GMT, but I meant EST) and came out with:

13 wins
7 losses
1 trade reversed with an opposite signal. if you would have exited, you would have lost 7 pips on the original trade, or 1/4 of the 2xATR value. If you would not have exited, it would have been a losing trade.

Had to give up 4 or 5 trades that were just outside the time rules, mostly around the 19:00-20:00 hour mark. IIRC the win/loss ratio was about the same, if not a little better.

Nobody is perfect, so there isn't really a way to know if these would have been actual trade results or not. Not to mention slippage/not getting TP orders filled when you normally think they would.

If I combine my july, august, and september visual backtest results, and discard the BE trades (anyone have a better idea of what to do with them to bring them into the analysis?), I get a record of 37 (W)-18 (L), or a win rate of 67%.

Is there a way to calculate expected account size if you trade fixed fractional MM and know your expectancy? You can't just multiply your expectancy by the number of trades can you? If I do that, I get different numbers than if I simulate a running total of my trades in a spreadsheet.

Can't you simulate approximate account balance with the following?

start balance = 100
balance after win = previous balance * 1.03 (3% risk/trade)
balance after loss = previous balance * 0.97 (again, 3% risk)
since multiplication is commutitive, order of trades does not matter

Inserted Code
Start    100
Win  103
W    106.09
W    109.27
W    112.55
W    115.93
W    119.41
W    122.99
W    126.68
W    130.48
Loss 126.56
L    122.77
L    119.08
L    115.51
End  115.51
Hindsight is 20/20
  • #618
  • Oct 11, 2009 1:28am Oct 11, 2009 1:28am
i have a dumb thought.

hope you guys help to clear it.

why not long both when you long the one doesn't make a lower low?

it seems to me all the successful trades you guys posted could be 2X more profitable if you trade both pairs.

i hope the profit is not due to recent ranging conditions.
  • #619
  • Oct 11, 2009 6:57am Oct 11, 2009 6:57am
No.

The key principle to this system is that you buy the relatively stronger pair or sell the relatively weaker pair. When a divergence occurs, what's really happening is the 2 pairs (which are normally correlated) are going out of correlation. So when the pairs go out of whack, you buy the suddenly stronger one or sell the suddenly weaker one. They will eventually revert to being correlated, but you'll be out of the trade by then.

To express it another way, if a pair makes a lower low then it's weak, right? Why would you buy it? But because the other pair failed to make a lower low it's stronger than the one that made the LL, so that's the one you'd buy logically.

The same logic applies to a sell decision. If a pair makes a higher high, then it's strong. So why would you sell it if it's strong? But the other pair failed to make a HH, so it's weak. Logically you'd sell the weak pair, not the strong one.
You are in a maze of twisty little passages, all alike.
  • #620
  • Oct 11, 2009 12:07pm Oct 11, 2009 12:07pm
Hi guys,

Just sifted through most of the thread and thought I'd get involved.

Probably an obvious answer but has anyone looked at taking ATR x3 for TP? I always feel more comfortable with systems that reward more than they risk. I expect the success rate would drop but would it drop enough to make it invalid?

Or trying 1.5x ATR for SL?
1 trader viewing now