The Bat...a strategy by Team Aphid | Page 24 | 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

  • #461
  • Feb 19, 2007 12:52am Feb 19, 2007 12:52am
We have tested 5 out of 8 currencies on your list except GBPCHF, AUDNZD and EURCAD. Since we are working on a few projects at the moment including improvement of BAT by using different exit strategies, we have no plans to backtest those three pairs with original rules. But we are more than happy to guide you if you do backtests yourself.



Quoting Gandolf
Hello, TA.

I have posted a collection of small charts that i use to keep track of volatility for all currency pairs. The 8 contained in this collection are the most volatile over a 30 hour period and are arranged in order of volatility. In that group is EURJPY which has not backtested well for you. Also, GBPCHF which you have not tested yet. Do you have any plans to back test it? I would be happy to do some grunt work for us all here but i am not as experienced so might need some guidance. Are there any others among this group that is on your hit list to test?
  • #462
  • Feb 19, 2007 12:53am Feb 19, 2007 12:53am
Quoting richy
Gandolf, could you please upload your ATR indicator? It's easier to monitor than the one provided in MT4. thanks
Hey, Richy. Here it is. PM me if you have questions about how to set it up.
Attached File(s)
File Type: mq4 Range MA.mq4   2 KB | 631 downloads
  • #463
  • Feb 19, 2007 12:56am Feb 19, 2007 12:56am
Quoting TeamAphid
We have tested 5 out of 8 currencies on your list except GBPCHF, AUDNZD and EURCAD. Since we are working on a few projects at the moment including improvement of BAT by using different exit strategies, we have no plans to backtest those three pairs with original rules. But we are more than happy to guide you if you do backtests yourself.
Have you posted anything already concerning this? Or would you like to provide that guidance here? or you can PM me if you prefer... whatever works best.

And... THANKS!
  • #464
  • Feb 19, 2007 1:03am Feb 19, 2007 1:03am
We think you are using a different server than Alpari according to the chart you attached. We sold @236.34 and still holding the trade with 300 pips profit and price has not hit Red BAT line yet, and red BAT line sits @ 233.6380 (rounded up to 233.64) where we will close our short position and get long with one lot.

Quoting Gandolf
I rode GBPJPY all the way from the entry, down past the 423.6 Fib level, and back up to the BAT trailing stop. It reversed as it approached the 232 psych level. Would have been nice to have saved a portion of that 146 pips before my exit.
  • #465
  • Feb 19, 2007 1:12am Feb 19, 2007 1:12am
Quoting TeamAphid
We think you are using a different server than Alpari according to the chart you attached. We sold @236.34 and still holding the trade with 300 pips profit and price has not hit Red BAT line yet, and red BAT line sits @ 233.6380 (rounded up to 233.64) where we will close our short position and get long with one lot.
Using FXDD. If i am going to backtest with your guidance, do i need to work from the Alpari data, or will FXDD be close enough?

Based on my FXDD charts i took a 300 pip profit and am now long in GBPJPY.

What about the 423.6 Fib level / major S/R level idea?
  • #466
  • Feb 19, 2007 2:34am Feb 19, 2007 2:34am
Hi nTeam A, please ahve a look the attached chart. I have included a Fractal indicator to help determine the different swing points. Once a new trade is triggered, would you uses Swing low A or B to draw the fib extensions. If I recall correctly you said the most recent low, which would be Swing low A .

The same applies to the low at 14 Feb and 13 Feb. The correct low would be 14 Feb
Regards Eckbert
Attached Image
  • #467
  • Feb 19, 2007 2:43am Feb 19, 2007 2:43am
All your thoughts, exits and filters may have a potential, but there is only one way to understand if they are tradable, backtesting. Therefore, you need to choose an exit strategy and start backtesting. Knowing the system better, gaining more confidence and having faith in the system; we believe that all will be obtained by backtesting.

We are ready to support you with all the help you need.


Quoting Gandolf
Someone else's idea that i have adopted as a mechanical exercise is updating support and resistance levels as follows.

- High and Low from last year and this year
- High and Low from last month and this month
- High and Low from last week and this week
- High and Low for prior trading day
- Psych levels at even 100 pip levels
- Oh, yeh... and also take note of the multiple time frame 20, 50, 100 and 200 Daily SMAs

What i am wondering is if it would be appropriate to exit when
- we pass the 423.6 Fib level
- price reverses at one of these major S/R levels
- moves away at a certain velocity
- and moves a certain distance

I rode GBPJPY all the way from the entry, down past the 423.6 Fib level, and back up to the BAT trailing stop. It reversed as it approached the 232 psych level. Would have been nice to have saved a portion of that 146 pips before my exit. For now, i'll give up the 146, but if TA thinks this could work I'd be happy to do the work to backtest this if someone can help with parameters.
  • #468
  • Feb 19, 2007 2:47am Feb 19, 2007 2:47am
For backtesting questions, please use PM to keep main thread focused for other contributers. Then, if we end up with a tradable result after backtesting, we can use main thread to open a discussion and get the ideas of other FF members.



Quoting Gandolf
Have you posted anything already concerning this? Or would you like to provide that guidance here? or you can PM me if you prefer... whatever works best.

And... THANKS!
  • #469
  • Feb 19, 2007 3:29am Feb 19, 2007 3:29am
Yes, you need to use Alpari server since all our tests are based on that server and we can compare the results.


Quoting Gandolf
Using FXDD. If i am going to backtest with your guidance, do i need to work from the Alpari data, or will FXDD be close enough?

Based on my FXDD charts i took a 300 pip profit and am now long in GBPJPY.

What about the 423.6 Fib level / major S/R level idea?
  • #470
  • Feb 19, 2007 3:40am Feb 19, 2007 3:40am
For the most recent GBPUSD long trade, we are using Feb 16 15:00 bar's low to draw fib. retracements. For Feb 13/14 lows, we used Feb 13 12:00 bar's low.

If you want to give more room to a trade and have larger stoploss, then you may choose above swing lows. But if you cannot risk that much stoploss or if you are conservative, then you may choose Feb 14 01:00 and Feb 16 22:00 bars as swing lows, but in our opinion, they are not significant. We prefer most recent low/high to be significant.



Quoting Eckbert
Hi nTeam A, please ahve a look the attached chart. I have included a Fractal indicator to help determine the different swing points. Once a new trade is triggered, would you uses Swing low A or B to draw the fib extensions. If I recall correctly you said the most recent low, which would be Swing low A .

The same applies to the low at 14 Feb and 13 Feb. The correct low would be 14 Feb
Regards Eckbert
  • #471
  • Edited 5:11pm Feb 19, 2007 5:33am | Edited 5:11pm
I add distance variable, for BAT line to be more flexible, it can be used to ignore spike.

If you're doing manual trade, you can put both into chart, I am using it to filter invalid signal which are :
1. Spike bar that crosses BAT line with its high/low is still in distance (11 pips)
2. Next bar after crossing bar cannot break high/low

Valid signal is if "crossing bar" break BAT line within distance and next bars beyond distance.

Attached Image
  • #472
  • Feb 19, 2007 5:18pm Feb 19, 2007 5:18pm
I took a little quick visual backtest and fix a little bug.

Attached Image


Attached Image


Attached Image
Attached File(s)
File Type: mq4 BAT ATR v1 distance.mq4   4 KB | 563 downloads
  • #473
  • Feb 19, 2007 6:16pm Feb 19, 2007 6:16pm
I think it would be a good idea if someone could make the BAT indicator to show the value of the BAT line for the current bar as well as the value of the BAT line for the last completed bar (maybe on the upper left corner?) It would save a lot of time! I would do it myself if I knew how to...
  • #474
  • Feb 19, 2007 6:43pm Feb 19, 2007 6:43pm
What you are doing is actually something very similar to change Factor parameter in original BAT indicator.

Please have a look at GBPJPY chart. Change distance parameter to 1.0 and keep ATR Period @ 5 and Factor @ 4.0

Then, in original BAT indicator, keep ATR period @ 5, but change Factor to 4.6. You'll see what we mean.

The result will be less noise, bigger stoplosses.

Thanks for your contributions.



Quoting mason
I add distance variable, for BAT line to be more flexible, it can be used to ignore spike.

If you're doing manual trade, you can put both into chart, I am using it to filter invalid signal which are :
1. Spike bar that crosses BAT line with its high/low is still in distance (11 pips)
2. Next bar after crossing bar cannot break high/low

Valid signal is if "crossing bar" break BAT line within distance and next bars beyond distance.

Attachment
  • #475
  • Feb 19, 2007 6:57pm Feb 19, 2007 6:57pm
has anyone tried this, treat each trade equally in terms of success probability. That is, first calculate number of pips needed for SL, and adjust the position size so that the risk for that trade is x%.

For instance, let say we enter 1 lot for a position of SL of 50 pips. Then if a trade is supposed to have a SL of 100 pips, we bet only 0.5 lots.
  • #476
  • Feb 20, 2007 2:06am Feb 20, 2007 2:06am
[quote=mason;222168]I took a little quick visual backtest and fix a little bug.

On the schedule 2 indicators are resulted: BAT ATR v1 and BAT ATR Distance (last version which the author has corrected from a mistake)
BUT! The mistake of new indicator BAT ATR Distance Is observed
EUR/USD concerns the remote indicator (11 pips) with a thorn and departs back, BUT! The indicator reacts (instead of ignores a thorn) and starts to draw from below a dark blue line. The arrow of white color specifies the new corrected indicator.
The previous indicator (first version BAT ATR Distance) was more correct. It ignored a contact of thorns and reacted only to the price of the closed position bar's. I do not understand for what you have altered it. The new indicator badly works, return please old version BAT ATR Distance
Attached Image
  • #477
  • Edited 6:21am Feb 20, 2007 6:02am | Edited 6:21am
Quoting TeamAphid
We would like to remind everyone that a better exit strategy is still needed
not to leave too many hard-earned pips on the table. That's what we are focusing as a team as well.
I have been looking into alternative exits for trade 1. One method that I use with another strategy is to tighten the stop (reduce ATR multiple) after hitting an initial profit target. Average profit on trade 1 is 111 pips, so tighten after say 200 pips. However, this just gets you out of the trend too early.

Looking at this, even the 4ATR stop gets out of the trending market too early in many cases (e.g. look at trades at end April-May). The strategy keeps reversing & taking losing trades against the trend. So I looked at widening the stop to 6 ATR after a gain of 200 pips (This would keep you in the end April trade until 8580, current exit 8215). This helped in some cases, and might be worth checking this out on an EA to optimise the 200 pip trigger & the 6ATR to see if there is a better combination. I think you might get a better result by leaving more pips on the table!

However, this analysis did trigger another thought. A big gain suggests that we have a trending market. Why not skip the contra-trend trade after a big gain? This did seem to improve the results overall. Considering trade 1 only (I didn't check limit orders), this approach would skip 14 trades, 9 of them losers & improve the overall reult by 219 pips (7% improvement for trade 1).
  • #478
  • Feb 20, 2007 6:25am Feb 20, 2007 6:25am
Thank you for different ideas. We devoted too much of our time to filter out some of the losing trades especially by trying different set of parameters, and we concluded that Atr period 5 with factor 4 is the best combination for our risk profile.

Regarding your assumption for trending markets, check out GBPJPY chart. It was on a clear uptrend since April 2006 and it may even continue despite recent pullback. Now, look at the short signal on Jan 3rd, 2007. It was a counter trend signal and it was triggered after a good gain (103 pips). So, if you pass this short signal, you'll miss 455 pips. The weakness of this strategy is that you may miss big pips during trend reversal and large retracement times.

Thanks for your contributions.



Quoting charlies
I have been looking into alternative exits for trade 1. One method that I use with another strategy is to tighten the stop (reduce ATR multiple) after hitting an initial profit target. Average profit on trade 1 is 111 pips, so tighten after say 200 pips. However, this just gets you out of the trend too early.

Looking at this, even the 4ATR stop gets out of the trending market too early in many cases (e.g. look at trades at end April-May). The strategy keeps reversing & taking losing trades against the trend. So I looked at widening the stop to 6 ATR after a gain of 200 pips. Although it helped in some cases, the result overall was negative. It might be worth checking this out on an EA to optimise the 200 pip trigger & the 6ATR to see if there is a better combination. I think you might get a better result by leaving more pips on the table!

However, this analysis did trigger another thought. A big gain suggests that we have a trending market. Why not skip the contra-trend trade after a big gain? This did seem to improve the results overall. Considering trade 1 only (I didn't check limit orders), this approach would skip 14 trades, 9 of them losers & improve the overall reult by 219 pips (7% improvement for trade 1).
  • #479
  • Feb 20, 2007 7:22am Feb 20, 2007 7:22am
Quoting TeamAphid
Thank you for different ideas. We devoted too much of our time to filter out some of the losing trades especially by trying different set of parameters, and we concluded that Atr period 5 with factor 4 is the best combination for our risk profile.

Regarding your assumption for trending markets, check out GBPJPY chart. It was on a clear uptrend since April 2006 and it may even continue despite recent pullback. Now, look at the short signal on Jan 3rd, 2007. It was a counter trend signal and it was triggered after a good gain (103 pips). So, if you pass this short signal, you'll miss 455 pips. The weakness of this strategy is that you may miss big pips during trend reversal and large retracement times.

Thanks for your contributions.
I might not have explained myself clearly. I agree that ATR factor 4 is best overall. However, I was proposing to change it only after a threshold is reached (say 200 pips gain). This way, you keep the initial risk the same, but give the market more room when trending (see chart below).

In a similar way, you would only avoid the contra trend trade after a large (say 200 pip) gain. As a result, the proposed change would have taken the short signal on 3 Jan. I only looked at this on GBPUSD, so don't know if the same rule / threshold would improve the results on GBPJPY. If you have a spreadsheet on GBPJPY, I could check this fairly quickly.
Attached Image
  • #480
  • Feb 20, 2007 7:39am Feb 20, 2007 7:39am
Quoting richy
has anyone tried this, treat each trade equally in terms of success probability. That is, first calculate number of pips needed for SL, and adjust the position size so that the risk for that trade is x%.

For instance, let say we enter 1 lot for a position of SL of 50 pips. Then if a trade is supposed to have a SL of 100 pips, we bet only 0.5 lots.
I backtested it for the BAT with different scenarios, but only for January to end of July 2006. Send me a PM if you are interested.
7 traders viewing now