The Bat...a strategy by Team Aphid | Page 17 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

My Bat from team aphid trades 82 replies

The BAT Method - The EA 186 replies

BAT system for 15 min chart 28 replies

BAT indicator 2 replies

The BAT mathematical formula for lines? 0 replies

Attachments: The Bat...a strategy by Team Aphid
Exit Attachments

The Bat...a strategy by Team Aphid

  • #321
  • Jan 25, 2007 9:14pm Jan 25, 2007 9:14pm
Sorry, I haven't made the time to test this idea, but I post it here, in case anybody wants to investigate further. I actually thought of this some time back, but was too embarrassed to post something that I hadn't tested properly, in case it turned out to be a dismal failure. And it seemed so simple that I figured it must have already been tested, and discarded.

My first take was that an uptrend is defined by progressively higher BAT buy signals, and higher sell signals; a downtrend by progressively lower ones.

Then I modified the idea as follows: enter long only if the current buy signal is higher than the most recent sell signal; enter short only if the current sell signal is lower than the most recent buy signal. Or, in other words, if the (Trade #1 portion of the) previous trade was (or would have been) a winner, don't immediately enter in the opposite direction. This tends to prevent trading against some of the more "severe" trends.

I had a quick look at GBPUSD,H1 and it seemed to filter out some losing Trade#1 trades. Obviously it would need to be tested more exhaustively, to see whether the winners filtered out (especially Trades #2, #3, #4) offset the losing Trade #1s. As far as I can tell, the situations where winning Trade #1s get filtered out are where:

  1. there is a significant enough retracement following a huge winning trade (then the retracement is missed);
  2. there is a deep enough retracement in the prevailing trend (then the next impulse move is missed);
  3. prices are moving in some kind of widening wedge.

An obvious variation would be to take smaller size positions (possibly for Trade #1only?) when trading contrary to the above rule. The trend determination could also allow re-weighting the sizes across trades #1, and #2-4, if the trend is particularly strong (e.g. three successive alternating signals are progressively higher, or lower).

David

  • #322
  • Jan 25, 2007 9:15pm Jan 25, 2007 9:15pm
Quoting TimeFreedom
I have been looking at my charts and think maybe we can reduce the number of false trades by placing the 1st buy/sell stop at the high of the bar that breaks through the BAT lines. Obviously you will need to wait for the bar to close.

TimeFreedom

Wow you noticed what I have, but I think a few pips above the high of the breakout candle.
  • #323
  • Jan 25, 2007 10:36pm Jan 25, 2007 10:36pm
Facinating thread. I've read it again from one to ThreeTwentyTwo and gained much in tips and insights. Especialy the various methods and ideas on filtering and trend indication.

I am still having good success utilizing the 15m chart with modifyed ATR BAT settings, nothing really to report. But it is proving to be helpful in validating entry/exit setups.

: ) Thom
  • #324
  • Jan 26, 2007 3:48am Jan 26, 2007 3:48am
Thank you for your suggestion, David. Please don't be embarrassed because the main goal of this thread is throwing ideas, testing them and seeing if they work.

We'll start testing your idea, but as you all know, it may take some time to finish it as we are using 1Hr charts.

Thanks again for all your contributions to this thread.


Quoting hanover
Sorry, I haven't made the time to test this idea, but I post it here, in case anybody wants to investigate further. I actually thought of this some time back, but was too embarrassed to post something that I hadn't tested properly, in case it turned out to be a dismal failure. And it seemed so simple that I figured it must have already been tested, and discarded.

My first take was that an uptrend is defined by progressively higher BAT buy signals, and higher sell signals; a downtrend by progressively lower ones.

Then I modified the idea as follows: enter long only if the current buy signal is higher than the most recent sell signal; enter short only if the current sell signal is lower than the most recent buy signal. Or, in other words, if the (Trade #1 portion of the) previous trade was (or would have been) a winner, don't immediately enter in the opposite direction. This tends to prevent trading against some of the more "severe" trends.

I had a quick look at GBPUSD,H1 and it seemed to filter out some losing Trade#1 trades. Obviously it would need to be tested more exhaustively, to see whether the winners filtered out (especially Trades #2, #3, #4) offset the losing Trade #1s. As far as I can tell, the situations where winning Trade #1s get filtered out are where:

  1. there is a significant enough retracement following a huge winning trade (then the retracement is missed);
  2. there is a deep enough retracement in the prevailing trend (then the next impulse move is missed);
  3. prices are moving in some kind of widening wedge.

An obvious variation would be to take smaller size positions (possibly for Trade #1only?) when trading contrary to the above rule. The trend determination could also allow re-weighting the sizes across trades #1, and #2-4, if the trend is particularly strong (e.g. three successive alternating signals are progressively higher, or lower).

David

  • #325
  • Jan 26, 2007 3:59am Jan 26, 2007 3:59am
Quoting hanover
Sorry, I haven't made the time to test this idea, but I post it here, in case anybody wants to investigate further. I actually thought of this some time back, but was too embarrassed to post something that I hadn't tested properly, in case it turned out to be a dismal failure. And it seemed so simple that I figured it must have already been tested, and discarded.



David
Hi David

A quick look on the chart shows that this techtique would have saved a number of false signals; so would have trades in the direction of the 72EMA which was mentioned by a member earlier.

The only drawback is that you may miss a potential reversal point, it really needs a proper forward testing (not back test) to prove which one, filter or no filter, is advantageous. Eye balling shows: Filter
Regards
  • #326
  • Jan 26, 2007 4:49am Jan 26, 2007 4:49am
Just for your info., we did a quick test (just for two months in 2006) by using David's following filter and the results are -58 pips for November 2006 and -428 pips for December 2006. The live trading for these months produced +82 and +520 pips for us. We'll do some forward testing as well.

Quoting hanover
Sorry, I haven't made the time to test this idea, but I post it here, in case anybody wants to investigate further. I actually thought of this some time back, but was too embarrassed to post something that I hadn't tested properly, in case it turned out to be a dismal failure. And it seemed so simple that I figured it must have already been tested, and discarded.

My first take was that an uptrend is defined by progressively higher BAT buy signals, and higher sell signals; a downtrend by progressively lower ones.

Then I modified the idea as follows: enter long only if the current buy signal is higher than the most recent sell signal; enter short only if the current sell signal is lower than the most recent buy signal. Or, in other words, if the (Trade #1 portion of the) previous trade was (or would have been) a winner, don't immediately enter in the opposite direction. This tends to prevent trading against some of the more "severe" trends.

I had a quick look at GBPUSD,H1 and it seemed to filter out some losing Trade#1 trades. Obviously it would need to be tested more exhaustively, to see whether the winners filtered out (especially Trades #2, #3, #4) offset the losing Trade #1s. As far as I can tell, the situations where winning Trade #1s get filtered out are where:

  1. there is a significant enough retracement following a huge winning trade (then the retracement is missed);
  2. there is a deep enough retracement in the prevailing trend (then the next impulse move is missed);
  3. prices are moving in some kind of widening wedge.

An obvious variation would be to take smaller size positions (possibly for Trade #1only?) when trading contrary to the above rule. The trend determination could also allow re-weighting the sizes across trades #1, and #2-4, if the trend is particularly strong (e.g. three successive alternating signals are progressively higher, or lower).

David

  • #327
  • Jan 26, 2007 6:23am Jan 26, 2007 6:23am
Short cable @ 1.9595 with limit orders in place. Just curious to know TA's thoughts on the swing high for this trade. I placed mine at 1.9681, which is also an hourly pivot level.
Attached Image
Michael fxgorilla
  • #328
  • Jan 26, 2007 6:55am Jan 26, 2007 6:55am
Unfortunately, at this time my ECN broker (efxgroup) does not offer a charting package or datafeed for metatrader. Therefore, I'm curious to how you would deal with the big difference in spread when using Alpari data with my ECN order entry. On the cable, it's no big deal and usually within a pip or 2, but for EURAUD, it's always much more. For example, right now Alpari is quoting EURAUD at 6690 ask and 6675 offer, while EFXGroup is quoting 6684 ask and 6679 bid. I find that most times, there is a pretty constant 5 pip difference between efx and alpari, with efx always inside alpari's spread. Therefore, would it make sense to use Alpari data, then adjust for the quote difference when placing stop or limit orders. For example, right now using Alpari data, we'd have a sell stop in place to short EURAUD at 6638. If I entered my order to sell at 6643, and did the same for all future orders, would that in the long run effectively compensate for the discrepency?
Michael fxgorilla
  • #329
  • Jan 26, 2007 7:04am Jan 26, 2007 7:04am
We placed our swing high @ 1.9679, high of 08:00hr bar (Alpari server), which is almost identical with yours.


Quoting fxgorilla
Short cable @ 1.9595 with limit orders in place. Just curious to know TA's thoughts on the swing high for this trade. I placed mine at 1.9681, which is also an hourly pivot level.
  • #330
  • Jan 26, 2007 7:10am Jan 26, 2007 7:10am
You are right and we are doing the same way you asked, adjusting the quote difference between Alpari and our broker when placing stop and limit orders.
We believe this method would compensate effectively in the long run.
Good point, thanks for asking.


Quoting fxgorilla
Unfortunately, at this time my ECN broker (efxgroup) does not offer a charting package or datafeed for metatrader. Therefore, I'm curious to how you would deal with the big difference in spread when using Alpari data with my ECN order entry. On the cable, it's no big deal and usually within a pip or 2, but for EURAUD, it's always much more. For example, right now Alpari is quoting EURAUD at 6690 ask and 6675 offer, while EFXGroup is quoting 6684 ask and 6679 bid. I find that most times, there is a pretty constant 5 pip difference between efx and alpari, with efx always inside alpari's spread. Therefore, would it make sense to use Alpari data, then adjust for the quote difference when placing stop or limit orders. For example, right now using Alpari data, we'd have a sell stop in place to short EURAUD at 6638. If I entered my order to sell at 6643, and did the same for all future orders, would that in the long run effectively compensate for the discrepency?
  • #331
  • Jan 26, 2007 7:14am Jan 26, 2007 7:14am
Thanks TA! One other question - with regards to news. For example, right now with news coming out that could dramatically affect the cable, do you leave your limit orders in place, or cancel and replace once the whipsaw has settled.

Quoting TeamAphid
You are right and we are doing the same way you asked, adjusting the quote difference between Alpari and our broker when placing stop and limit orders.
We believe this method would compensate effectively in the long run.
Good point, thanks for asking.
Michael fxgorilla
  • #332
  • Jan 26, 2007 7:23am Jan 26, 2007 7:23am
Team Aphid,
I like your indicator but I don't know Metatrader language. If possible,I would like to know the formula with common math lanuguage so I can see it on tradestation and see how it works on stocks and futures. When you want

regards
  • #333
  • Jan 26, 2007 7:53am Jan 26, 2007 7:53am
We have forwarded your request to our programmer.


Quoting canados
Team Aphid,
I like your indicator but I don't know Metatrader language. If possible,I would like to know the formula with common math lanuguage so I can see it on tradestation and see how it works on stocks and futures. When you want

regards
  • #334
  • Jan 26, 2007 9:12am Jan 26, 2007 9:12am
Quoting TeamAphid
We have forwarded your request to our programmer.
thank you very very much
  • #335
  • Jan 26, 2007 9:48am Jan 26, 2007 9:48am
Quoting canados
Team Aphid,
I like your indicator but I don't know Metatrader language. If possible,I would like to know the formula with common math lanuguage so I can see it on tradestation and see how it works on stocks and futures. When you want

regards
I am also interested in reproducing this indicator for FibonacciTrader v4.0!
  • #336
  • Jan 26, 2007 11:58am Jan 26, 2007 11:58am
Thank you TeamAphid for nice system.

I've been using your indicator for a while (PF) to trail stops and like it very much, many thanks.
I also just started experimenting, using it for entries with a small adjustment. Instead of taking initial position on price touching 'bat', I'd wait for 50% fib ret with some kind of confluence and open 2 units.
Attached Image
  • #337
  • Jan 27, 2007 12:22am Jan 27, 2007 12:22am
Quoting TeamAphid
Just for your info., we did a quick test (just for two months in 2006) by using David's following filter and the results are -58 pips for November 2006 and -428 pips for December 2006. The live trading for these months produced +82 and +520 pips for us. We'll do some forward testing as well.
Interesting. Of course I don't have the trade-by-trade details of your test results, but I suspect this means that some of the biggest moves are being filtered out (if I understand correctly, BAT trade #1s are a trend following system, which generally require a few sizeable moves to offset the more frequent small losses). Of course the filtering will also be eliminating trades #2-#4 whenever it occurs. Hence my subsequent idea for re-weighting, rather than removing the entire set of trades. Currently the prospective "ranging component" is out-weighing the prospective "trending component" in each set, at least potentially, by 3 units to 1.

Elliott wave theory tells us that wave 3 is generally the longest impulse wave. I would have thought that the filtering would be likely to preclude entry on wave 1 than wave 3, unless the wave 2 retracement is significant. Hopefully the attached chart clarifies my point. The trend filter eliminates the marginally profitable wave 1 trade (red X), and also the losing wave 2 trade (blue X), but permits entry into the hugely profitable wave 3 trade (which actually gets stopped out as it hits the prior swing low, but that is an exit issue). The Elliott wave truism was a consideration behind my idea.

Thanks for testing this. Hopefully, a post-mortem along the lines of (i) separating results for trade #1, and trade #2-4, and (ii) whether or not it's the fact that a few sizeable wins are being missed, and where these big wins occur during a trend (early, midstream, late) etc, will tell you why applying a trend filter has proven unprofitable. It seems odd to me that a predominatly trend following system would not be enhanced by some kind of trend filtering.

David
Attached Image
  • #338
  • Jan 27, 2007 1:16am Jan 27, 2007 1:16am
We trade BAT 100% mechanical, therefore, we leave our limit orders in place during news time.


Quoting fxgorilla
Thanks TA! One other question - with regards to news. For example, right now with news coming out that could dramatically affect the cable, do you leave your limit orders in place, or cancel and replace once the whipsaw has settled.
  • #339
  • Jan 27, 2007 1:53am Jan 27, 2007 1:53am
Thanks for contributing. You can have tighter stop loss with this method but you can miss out some very profitable trades if the breakout is strong. Check out GBPJPY long trade on Jan 11th, 2007 @07:00 bar. You miss 859 pips profit if you use this method for entry. Or check out GBPUSD long trade on Nov 17th, 2006 @ 16:00 bar. You miss 533 pips profit. It is essential to catch these big movers.

Your idea should be tested properly, but we believe the amount of pips you save will not compensate the ones from miss-out trades.



Quoting raczekfx
Thank you TeamAphid for nice system.

I've been using your indicator for a while (PF) to trail stops and like it very much, many thanks.
I also just started experimenting, using it for entries with a small adjustment. Instead of taking initial position on price touching 'bat', I'd wait for 50% fib ret with some kind of confluence and open 2 units.
  • #340
  • Jan 27, 2007 2:02am Jan 27, 2007 2:02am
Hi David,

We have tested your following idea only, not the one "re-weighting the sizes accross trade # 1". Re-weighting would be safer not to miss out any trades.

On the other hand, we haven't thought about using Elliot Waves because we believe the waves add more subjectivity to mechanical systems.

Thanks for your time and valuable contributions.



Quoting hanover
Then I modified the idea as follows: enter long only if the current buy signal is higher than the most recent sell signal; enter short only if the current sell signal is lower than the most recent buy signal. Or, in other words, if the (Trade #1 portion of the) previous trade was (or would have been) a winner, don't immediately enter in the opposite direction. This tends to prevent trading against some of the more "severe" trends.


David
1 trader viewing now