[costard]

SOME STATS ON ASH’s (thanks again Ash) AMAZING SYSTEM!

I’ve been doing some basic analysis. The process isn’t finished by any means but I wanted to get something up in the interim for debate/observation/criticism.

Here’s what I have done. I manually traded the daily EUR/USD from January 2006 to now using FXDD platform. There were 56 valid trade Trigger Day Candles according to Swingman’s neat Indicator (thanks Swingman). Entry and exits were as Ash describes in the pdf on page 1.

There was no filtering, no MAs or CCIs. It was basic Ash rules with TP1 taken at first profitable candle, moving stop to BE and taking TP2 at AO colour change. The Initial Stop Loss (ISL) was at the Swing High /Swing Low.

I traded two units per trade worth £100 each - £200 representing 2% of a £10,000 account. The financial risk on each trade was therefore the same with the number of lots calculated by dividing £200 by the ISL in pips and then splitting into two units. e.g. Initial SL of 300 pips. 200/300 = 0.66 therefore each unit is 0.33 lots.

RESULT
Over the entire period there were 30 winning trades and 26 losing trades and a financial LOSS of some £1,990. I then decided to add some filters!

1) First I discounted any trades where the direction of Ash’s arrow wasn’t confirmed by the trend of the previous week on the Daily Chart (note NOT the trend on a Weekly Chart). To determine this I ran the cross hairs from the Close of the 5th candle back to the Close of the Trigger candle. This gave me a trend direction. If the trend was in the opposite direction to the arrow the trade was rejected.

RESULT
Doing this excluded 6 of the trades. Now there were 28 wins, 22 losers and the LOSS was now only around £1,400.

2) As has been mentioned here in previous posts by several people I next considered the size of the Initial Stop Loss (ISL).

A large ISL inevitably increases the win/loss ratio but it can severely curtail profits. I tested different ISLs based upon Fibonacci ratios of the Swing High/Swing Low and similarly using the length of the Trigger Day candle itself (from L to C for UP candles and from H to C for DOWN candles).

Here are the results (rounded):

# of trades = 50

Initial SL WIN LOSS P/L

Full Swing 28 22 -£1,400
62% 28 22 £450
50% 26 24 £1,090
38% 21 29 £950
23% 17 33 £2,560

Trig. Candle 26 24 £1,080
62% 20 30 £2,530
50% 17 33 £890
38% 16 34 £2,980
23% 13 37 £6,550

RESULT
Tempting though it is to draw conclusions from these figures 50 trades are far too few to be certain from a statistical point of view - although I tried to maximise the validity of the results by splitting the data in half and then confirming the findings of the first 25 with the results from the second batch.

Nevertheless it would seem to be reasonable to say that using an ISL of the Full Swing is not going to be the most profitable choice. At the same time the wholly unrealistic 65.50 ROI with the 23% Fib of the Trigger Candle cannot be relied upon being generated from evidence of only 13 winning trades.

Of the other fib ratios the most consistent was the 38% Fib of the Trigger Candle. It may be a little more realistic but this is still only based on 16 winning examples where just 5 of them provided 75% of the total profit. But this is not unsurprising. Indeed it is to be expected because as far as I can see an inherent aspect of the trading style of Ash’s system is that the “serious money” is made from relatively few trades which catch a strong trend.

The other drawback, of course, is that with a small SL you have to endure considerably more losing trades. Nevertheless it seems on the face of it that to make money with Ash’s system there is logic in choosing the smallest ISL your temperament can stand to ensure that when it finds a trend you are in deep enough for a good reward. The old “cut the losers short and let the profits run mantra”.

Also bear in mind that even the default position produced 22 losing trades out of 50.

3) I wondered also whether there was any significant relationship between the success of a trade and the robustness of the trend immediately preceding it. I was already filtering to ensure the 5day Trend was going the right way.

Therefore I added another filter to compare the size of the 5day Trend in pips in relation to the 10 Day Average Trade Range - the theory being that triggers with stronger positive trends behind them were more likely to be more successful. e.g (5day Trend)/(ATR(10)) x 100%

RESULT
Once again the amount of data is not sufficient to draw great significance.

One observation worth making though is that profits started to tail off heavily once the 5day Trend exceeded 100% of ATR(10) – presumably an indication that the trend had already run out of steam.

Also filtering with the Trend/ATR percentage had a measurable effect upon the win/loss ratio even though the profits remained pretty consistent across the range 0 to 100% - approx £3,000 to £4,500 using the 38% fib ISL and between £6,500 and £8,000 with the 23% fib ISL.

Here are Win/Loss percentages for the two SL values at different Trend/ATR % filter levels.

(Filter Levels) None 25% 45% 80% 100%

38% Trig Candle 32% 36% 41% 43% 48%
23% Trig Candle 26% 30% 32% 37% 43%

CONCLUSION

Clearly a lot more data is needed and I will endeavour to add to it as time allows. Once the evidence is statistically robust enough I will of make the full data available here.

Remember this only applies to EUR/USD. I have already started doing a similar exercise with GBP/USD which has very different characteristics.

I look forward to reading responses.