My Mechanical Trading System by VTyro

[Anonymous]

My very first post to this new forum and already in the title I'm fibbing!

Sadly I have only part of a trading system at the moment. Here's where I'm at...

I'm very interested in:

1. Trading with the trend.
2. Using price action to determine entry timing.
3. Cutting losses quickly and letting profits run.
4. Using money management strategies to minimize risk/maximize return.

OK, nothing revolutionary there...

In terms of point 1, I tried a number of different methods of assessing whether there is a trend - Effeciency Ratio; ADX; Chande Trend Index; Trend Detection Index etc. None of them quite did the job in the way I was using them. My suspicion is that these measures work better in leveraged markets such as futures but I've done no research on this. Finally, I found a method that gave me what I was looking for. This is a custom indicator based on the Guppy Multiple Moving Average approach.

For point 2, I've simply used a new 21 day high to enter. I'm not 100% about this as there may well be a 'cheaper' entry point than.

As confirmation, I've used a volume oscillator using the 5 days average volume and the 21 day average volume. I also have a filter for liquidity. That's it really for the entry side.

I then ran these criteria through tradesim using a timed stop as the exit - by the way, there appears to be a bug with this as if I specify an exit after 1 day the exit will be after 21 days. If I actually put in a 21 day exit, the exit will happen after 154 days...I believe the calculation is 20 + X * 7, where X is the number put into the code.

Anyway, the idea behind this test was to see if my entry method was better then random entry. In this case, I was expecting to see that winning trades exceeded losing trades and that the size of wins and the size of losses were about the same. Interestingly, the opposite happened over all timeframes - the percentage of winning and losing trades were about the same while the size of winning trades exceeded losing trades by a statistically significant amount. Still, this is a statisitcal edge so I'll take it.

The detail for each of the time frames is as follows:


GUPPY METHOD 21 DAY TERM
Trades Taken: 2230
Winning Trades: 956 (42.87%)
Losing Trades: 957 (42.91%)
Breakeven Trades: 317 (14.22%)

Largest Winning Trade/(Date): $8,107.65 (18-02-00)
Largest Losing Trade/(Date): -$4,870.84 (08-05-98)
Average Winning Trade: $588.99
Average Losing Trade: -$489.53
Average Win/Average Loss: 1.2032

GUPPY METHOD 28 DAY TERM
Trades Taken: 2040
Winning Trades: 904 (44.31%)
Losing Trades: 924 (45.29%)
Breakeven Trades: 212 (10.39%)

Largest Winning Trade/(Date): $411,021.45 (13-10-00)
Largest Losing Trade/(Date): -$4,942.53 (01-09-00)
Average Winning Trade: $1,136.51
Average Losing Trade: -$572.00
Average Win/Average Loss: 1.9869

GUPPY METHOD 35 DAY TERM
Trades Taken: 1953
Winning Trades: 891 (45.62%)
Losing Trades: 890 (45.57%)
Breakeven Trades: 172 (8.81%)

Largest Winning Trade/(Date): $16,621.96 (08-03-96)
Largest Losing Trade/(Date): -$4,870.84 (22-05-98)
Average Winning Trade: $783.80
Average Losing Trade: -$608.30
Average Win/Average Loss: 1.2885

GUPPY METHOD 42 DAY TERM
Trades Taken: 1885
Winning Trades: 873 (46.31%)
Losing Trades: 865 (45.89%)
Breakeven Trades: 147 (7.80%)

Largest Winning Trade/(Date): $12,368.27 (21-03-97)
Largest Losing Trade/(Date): -$4,932.66 (15-09-00)
Average Winning Trade: $822.64
Average Losing Trade: -$664.16
Average Win/Average Loss: 1.2386

GUPPY METHOD 49 DAY TERM
Trades Taken: 1836
Winning Trades: 898 (48.91%)
Losing Trades: 814 (44.34%)
Breakeven Trades: 124 (6.75%)

Largest Winning Trade/(Date): $15,202.01 (06-08-99)
Largest Losing Trade/(Date): -$4,873.07 (28-06-96)
Average Winning Trade: $870.96
Average Losing Trade: -$722.50
Average Win/Average Loss: 1.2055

GUPPY METHOD 56 DAY TERM
Trades Taken: 1783
Winning Trades: 855 (47.95%)
Losing Trades: 810 (45.43%)
Breakeven Trades: 118 (6.62%)

Largest Winning Trade/(Date): $12,828.55 (03-12-99)
Largest Losing Trade/(Date): -$4,879.53 (05-07-96)
Average Winning Trade: $950.91
Average Losing Trade: -$759.12
Average Win/Average Loss: 1.2526

GUPPY METHOD 63 DAY TERM
Trades Taken: 1745
Winning Trades: 838 (48.02%)
Losing Trades: 804 (46.07%)
Breakeven Trades: 103 (5.90%)

Largest Winning Trade/(Date): $22,500.00 (23-07-99)
Largest Losing Trade/(Date): -$4,885.94 (12-07-96)
Average Winning Trade: $1,039.93
Average Losing Trade: -$793.57
Average Win/Average Loss: 1.3104

GUPPY METHOD 70 DAY TERM
Trades Taken: 1707
Winning Trades: 837 (49.03%)
Losing Trades: 771 (45.17%)
Breakeven Trades: 99 (5.80%)

Largest Winning Trade/(Date): $18,750.00 (30-07-99)
Largest Losing Trade/(Date): -$4,883.81 (19-07-96)
Average Winning Trade: $1,073.29
Average Losing Trade: -$840.29
Average Win/Average Loss: 1.2773

I'm now up to point 3, which is essentially designing exit and stop loss criteria. One of the things to consider on this is whether the data above suggests having some time based stop built into the system itself. A further question is if so, would it be different for losing trades as opposed to winning trades? I suspect it should as both losing trades and winning trades get bigger the longer they're open. I'm considering some approach like "exit the trade if it's showing a loss after X days".

I'd be grateful for any thougths, suggestions, criticisms anyone might have.

Tyro

[Jade80121]

I was reading what you had to say,

Reguarding the 21 day high try this....
Plot a Donchian channel of 21, I personally do not like to use highs because I think they give to many head fakes (I plot Closes). Then plot a larger Donchian channel say a 144

I think taking the 21 day high is a great thing but you have to make sure that trend is favorable in your direction because 21 is still a small number compared to the larger overall trend? Second I think if the overall trend is up (the larger donchian channel moving up) then what I would wait for is for the 21dc to move DOWN and then as soon as it starts to pop UP on that very FIRST move up that would be a buy signal.... Also leaving maybe an ATR Stop of some sort to get out in case of trouble.

So in short buying a pullback of a stock moving up with lower risk due to waiting for a pullback. (or going short in the opposite)


Heck I have not tried it so I don't know but just a thought.....

[aukc]

Hi VTyro,

How do you code your custom indicator based Guppy MMA. I thought Guppy MMA is more of a visual tool? I am new to Metastock coding, sorry if I ask a stupid question.

Regards,
KC