The Bat...a strategy by Team Aphid

Quoting fukinagashi

Disliked

Team Aphid.

thanks for this method. I think it has potential and I like the fact that it is different. It reminds me very much of all the old PSAR systems, but I like the way you integrate Fibs for SL/TP and averaging.

I am nevertheless professional sceptic and so I am in the process of finalizing an EA for this system. It is intended to basically serve for backtesting and maybe limited fronttesting, to see how this system really behaves in ranging markets and with other currencies.

Would you guys rather put me that EA in a different thread, so that a discussion of the EA does not interfere with the discussion over the system, or do you prefer, to keep it all together?

Ignored

Quoting baddison

Disliked

Team Aphid, is there a reason why you don't use the 50% fib retrace as well?

Ignored

Quoting Swami S

Disliked

Team Aphid ,
May I just join the cheers of thankyou for this very clean system as yet i have no questions other than why did you call it the Bat?. While studying and testing in the pf I had nicknamed it the special love indicator so taken with it i was,
And now you have shared the mystery. I love the name . Please tell
Best Wishes
Swami S

Ignored

Quoting RandyG67

Disliked

According to the .pdf description, single lot trades are closed at the break of the next colored BAT bar. In this example, some good profits would be given back on the LONG trade thru the break of the Red BAT bar. Is this simply part of this strategy or am I misinterpreting?

thanks,
Randy

Ignored

Quoting baddison

Disliked

Team Aphid, instead of placing a stop loss to exit Entry Trade #1 at the red line (...or blue...), can you instead place a C.A.R. order? Will it have the same effect as stop one order, and enter one new order?

Ignored

Quoting fukinagashi

Disliked

I do not want to speak for them, but there is no 50% Fib. Fibonacci never spoke of the factor 0.5 and this figure is as well not derivable of his system. I nevertheless agree with you that a 50% retracement is not uncommon.

I am though curious why they "ignore" the fibonacci factor 0.236/1.236.

Ignored

Quoting hanover

Disliked

TA asked for feedback. Here are my initial thoughts. These are very general in their nature; I haven't done any testing.

I like BAT’s simplicity and precisely defined nature, as it allows for objective usage and automated testing.

IMHO the basic premise is very sound, in that (with trade #1) one is entering a newly emerging trend, while trades #2-4 seek to exploit early retracements in this trend, via secondary entries at a more favorable price. Re timing of entry, some systems seek to enter earlier (for a long position), as a pullback in the prevailing trend is falling back, or has fallen back, into an “oversold” area of prior support, thereby anticipating the bounce, while others enter later, giving the newly emerging move more time to validate itself. I would suggest that BAT’s trade #1 fits the latter category, and is therefore well-suited to a trending market. You are right.

Trades #2-4 also act as a kind of hedge, taking profit if the best the market manages to achieve is movement back to the starting point (100.0 line). This is perhaps significant, because these are the only profit-taking points, and any trailing stop method (on trade #1 in BAT’s case) always forfeits profit while waiting for the final retracement to hit the stop. Given that BAT has no explicit “move stoploss to breakeven” rule (not necessarily a bad thing, IMHO some traders do this too prematurely), one must come to terms psychologically with the fact that a trailing stop system occasionally allows a profit to erode itself into a loss (see RandyG67’s post).

For better or worse, with BAT one is constantly “in the market”, as the trailing stop for the current phase simultaneously heralds the entry, in the opposite direction, into the next phase. There is no definition of indeterminacy, nor is there a longer term trend indicator that restricts trading against the prevailing trend. True.

As I see it, one unnecessarily risky scenario is where the trailing stop (BAT line) lies just outside the 61.8 and (especially) the 38.2 lines, where new trades are being entered, as a small amount of volatility will cause these trades to be stopped out, before they have a chance to blossom. If it is too close, like 2-5 pips, you may not enter those trades. We agree. Perhaps the system could benefit from a variation along the lines of trades #2-4 only being entered after price, having retraced to the appropriate Fibo level, has subsequently moved X pips clear of the BAT line. Re volatility, the H1 chart is sufficiently short time frame to be buffeted by news announcements, and the spread costs are relatively higher to profit potential, compared with longer time frames (H4, D1, etc).

I note that TA allows a second lot to be opened initially (potentially making 5 trades). Extending this idea, I would suggest that in a “trending” market, it is preferable to weight the initial trade, by adding lot(s) at this point, since, the stronger the trend, (1) the less likely deep early retracements are to occur, and (2) in a strongly trending market, one wants to have as many lots on as possible (some traders even go as far as pyramiding), because this is where the longest moves, and hence high-pip profits, are most likely to eventuate. We tested this, because "first lot only" trades are profitable. You can have 2nd lot, not only for more profit, it is also advisable since it compensates spread cost.

Conversely, in a “ranging” market (in an H1 context), I would look to weight trades #2-4 more heavily, because trade #1 becomes a more risky proposition, in that (as a later type entry – see my earlier comment) it is more likely to lie nearer to the “wrong” extreme of the range.

Of course all of this presupposes that we’re somehow able to anticipate when the market is going to move in and out of trending and ranging phases, which is not necessarily an easy task! Difficult one.

The other key variable that could be experimented with is the “Factor”, but in attempting to optimize this, one must consider the caveats involved with “curve fitting”. (I note that the Factor does not have be an integer value.) Anyway, for whatever it’s worth, increasing the Factor allows a wider trailing stop, giving the trades more initial breathing space, but the drawbacks being that (1) losses will be greater, in a losing situation; (2) more profit is being forfeited when the BAT line finally gets hit, in a winning situation, and (3) this also means that the subsequent trade, on the phase change, is entered later (i.e. on balance, at a less favorable entry price). Decreasing the Factor would achieve the converse of all this. Attempting to optimize any system invariably results in these types of compromise, i.e. one can’t have one’s cake and eat it too! To be honest, we played with factor to see how it affects trades and also tested the same factor (factor 4) for other currencies and we believe factor 4 is suitable. But you can use the one more suitable to your risk profile.

Having not done any testing, the following is just a gut feel, but I might look to use a lower Factor value in a trending market, where one can arguably get away with setting tighter stops, and a higher Factor value in a ranging market, where positions stopped out prematurely have a higher probability of bouncing back.

However, and again a gut feel, my preference would lie with re-weighting the entries, rather than tinkering with the Factor.

RandyG67’s post underlines the possibility of taking some profit, perhaps a partial closure at the 161.8 line? We tested this and following BAT line for trade #1 exit is more profitable. Especially, in trending markets, 161.8 is too close, even for partial closure. Please see Nov 17th, 2006 - Nov 29th, 2006 trade. Of course this would reduce overall profit when the market happens to subsequently trend in prolonged fashion; another delicate compromise as we are decreasing both potential risk and return. So true. This also revisits the possibility of a lower Factor which would tighten the BAT line stop, locking in profit more quickly. Or even (but this would undermine the simplicity of the system) a twin-stoploss system, with half of the profit being saved by a “move secondary stop to breakeven” at some tactical point, and the other half allowed to “ride” on the BAT line. Good idea, worth testing.

This is probably only of academic interest, but increasing the “ATRperiod” value means that a greater number of historical bars are being averaged, generally resulting in a smoother moving BAT line (but weighting its value less heavily on the most recent data).

A couple of parting thoughts. One final addition would perhaps be some kind of longer term trend indicator, with the proviso that one trades only in the direction of this prevailing trend. We used many techniques to find this special trend indicator and still searching for it. And perhaps some special conditions involving the handling of potentially high-impact news announcements (but this would mean moving away from a mechanical, technically-based system).

Very good ideas and observations, please keep contributing. Thanks for your time.
Hope this gives some ideas to ponder. Comments and criticism most welcome!

David

Ignored

Quoting Lou G

Disliked

Perhaps someone will clarify the entry point for me - the bar (or candle) must actually make contact with the blue or red line (not pass through the point where the line used to be), is that correct?

I'm just trying to do a little eyeball backtest and am seeing a lot of places where price just misses the line by an hour or two and makes me wonder...

Also, is actual contact with the line necessary for exit as well?

Thanks,
Lou G

To add an example, is the crossing of the LEVEL of the blue line on 2007-1-16 at 15:00 a failed short trade that is stopped out when price crosses the subsequent red line?

Ignored

Quoting accrete

Disliked

OK, i'm treading lightly and humbly here.

I'll delete this post if it is not appropriate (or it can be moved to a BAT-Mod thread?) though it appears in reading this thread several times that we are looking for ideas/refinements...so posted below i share the chart of my trades during the 17th's London Open.

As you can see it is the dreaded 15m chart...! But it's become one of my favorite charts as i'm in front of my computer from 06:00 GMT to 10:00 GMT and then set chart alarms with a wake up call at 12:00 GMT and watch the charts for another 2 hours.

This is not purely mechanical, but my goal is to find as many reasons NOT to trade as i can. So by having the Channels and the dialy moving average "Wave" as filters i can hopefully catch the trend of the London open and go with the flow. Remember, my style is to get in and out as i'm still not comfortable leaving trades unattended. The BAT ATR is adding a measure of confidence that i hadn't had before. So if nothing else TeamAphid's thoughts shared in this thread have increased this trader's level of confidence in his personalized style. . . And That Is A Good Thing . . . even if what i'm sharing is now considered a "Hybrid". (Scott, your post is great and i agree 100% with your statement!)

OK, so here is the chart. I took long entries at each green circle and got out when the top of the channel was hit. Yes these are three seperate trades. The average hight of the channel was about 30 pips and these trades were worth about 20 to 25 pips each (21 to 27 to be exact). I could have just let this ride and not had the additional spread costs, but a pip in the bank is a good thing (of course this was in demo, but solid trades per my rules/style). These trades increased my demo account by just a smidge over 5% today.

Please don't shoot the messenger... I'm just sharing what worked for me and how greatful i am for this additional "filter" in my arsenal...and in that spirit i share it w/U all : )

Cheers,
Thom

http://www.accrete.com/fx-bat/charts/20070117a.gif

Ignored

Quoting Barcode

Disliked

TA
I am a little confused by your chart describing the Bat Range Trade on page 8 of the pdf document.

The pink arrows show the first 2 trades. You have the first pink arrow at the 61.8 line and the second at the 'Enter 1 lot' line. Shouldn't these pink arrows be reversed as the first trade should be at the 'Enter 1 lot'.

I presume price must break thru the fib level and not just touch it otherwise the bar after the first aqua arrow would have been a trade at the 38.2 level.

What constitutes a retrace? Does the price have to close inside the fib level and then break it for a valid retrace? Is the last bar on the chart a valid retrace and should there be a trade at the 61.8 level?

Ignored

Quoting TeamAphid

Disliked

Hi Barcode

Sorry you find this chart confusing but we were trying to point out the retracement trade pickup points not really the initial entry but I can see how this could confuse.

The first bar after the first aqua arrow may well have picked up the trade but as we all know it might not have as well so for the purpose of clarity and visual reference the definite touch of the 38.2 level 5 bars later was used.

The last bar of the chart is not a retracement as price had not returned to the initial entry level.

A retracement re entry has to occur only after price has returned to the initial 'Enter 1 lot' line as it did 4/12 at approx 12.00 bar. We are tring to catch trades as they range before hopefully heading in our desired direction.

Watching this type of trade and strategy in real time will help immensely.

Hope this helps clear some confusion

Ignored

Quoting Gmak

Disliked

just finished reading this thread and just about to start testing your sytem.

One point I'd like to make so far is that within your system there seem to be some pretty spectacular drawdowns that aren't highlighted among your monthly figures

I imagine you already know this but I'd like to draw your attention to two periods.

May 4th - May 11th - drawdown = -842 pips

Oct 31st - Nov 17th - drawdown = -920 pips.

Ignored

Quoting Gmak

Disliked

just finished reading this thread and just about to start testing your sytem.

One point I'd like to make so far is that within your system there seem to be some pretty spectacular drawdowns that aren't highlighted among your monthly figures

I imagine you already know this but I'd like to draw your attention to two periods.

May 4th - May 11th - drawdown = -842 pips

Oct 31st - Nov 17th - drawdown = -920 pips.

Ignored

Quoting TeamAphid

Disliked

And some spectacular profits:

June 14th - Jul 3rd -profit = +1112 pips

Jul 14th - Jul 26th - profit = +906 pips

Sep 12th - Oct 4th - profit = +779 pips

We already know this drawdown issue, but it depends on how you look at the system and the risk profile of trader using it. We are not scalpers and most of the systems we use are medium to long term systems. Therefore, we are more interested in monthly figures.

Thank you for highlighting this issue.

Ignored

Quoting Mr Trend

Disliked

And you hit it on the money right there. Would you want to go through that, mentally? I wouldn't. That's a horrendous drawdown.

That kind of drawdown, when I'm developing a system, is far from my acceptable limits.

Ignored

Quoting TeamAphid

Disliked

Please see post #176.

We believe the ones who go through that, mentally, gain big pips. Trading psychology is the most difficult part in trading.

Most portfolio managers or money managers are evaluated with respect to their annual performances, not even monthly.

We are not saying drawdown is not an important issue and we believe a smooth and consistent yield curve is better and preferable than the one with big zigzags. But, unfortunately, we are having this kind of drawdowns occasionally and this is the nature of trend following systems as they use larger stoploss levels compared to other systems.

Ignored