GDX Trading Systems

At last I have completed enough research and back testing that I can now begin to share some findings regarding my self-appointed mission to create an automated trading system that is both reliable and very profitable. 

I'm on a journey and quite sure the path is without end, but that keeps it interesting, right?

The good news is that I am finding strategies that generally match my objectives. The bad news is that it takes eons of time to craft a polished finished product (strategy) and even then, the product could always be better.....somehow. There are always more 'what-if' questions left unanswered when designing trading strategies than any mortal will ever have time to figure out. 

But I guess the fact that one does not have to be correct all the time when trading the stock market to make money is a consolation. Actually, it's a BIG consolation, now that I think about it! One just needs the odds on their side and the descipline to execute a sound strategy to succeed.

Anyway, let's take a look at the findings I have to share for today.

The table below provides an overview of 5 trading strategies I created for trading the Market Vector Gold Miner ETF (GDX) using the daily time frame. The Think or Swim platform was used to design these strategies and provide a detailed back test of their performance since 2007. 

The buy/sell signals are determined at the end of the NYSE trading session for use at the open on the following trading day. 

The top portion of the table details the five trading strategies (Pivot, Gap, PercentB, Step and UlcerX) and their performance. Each trade used 100 shares. The total number of round trip trades was 319, of which an average of 80.56% were winners. The gross gain in total was $30,924 and the average gain per trade was $96.94.

The strategies vary in the number of trades they generated over the past 5 years (32 - 120), the winning percentage also varied considerably (56% - 97%) and the average gain per trade was remarkably wide ($61 - $192). 

But for what it's worth, I imagine a diversity of strategies - some highly profitable mixed with some highly dependable - isn't an altogether bad way to go.

GDX – Market Vectors Gold Miner ETF Individual LONG Strategy Results (2007 – 2012)
Strategy		Trades	Win/Loss	Win/Loss %	Total Gain	Avg Gain per Trade
1	Pivot	120	100W / 20L	83.3%	$8,454	$70.70
2	Gap	88	70W /  18L	79.5%	$5,341	$60.69
3	PercentB	33	32W /  1L	97.0%	$5,070	$153.63
4	Step	46	37W /  9L	80.4%	$6,891	$149.80
5	UlcerX	32	18W /  14L	56.2%	$6,160	$192.50
Total		319	257W / 62L	Avg 80.56%	$30,924	Avg $96.94
Strategy Indicators Used
Strategy Name
1	Pivot	Uses a % band above/below the Pivot Point Moving Average
2	Gap	Uses gap lower in price and a moving average of price
3	PercentB	Uses Bollinger Band PercentB movement within a range
4	Step	Uses Average True Range and the True Strength Index
5	UlcerX	Uses Ulcer Index and the True Strength Index
GDX – Market Vectors Gold Miner ETF Cumulative LONG Strategy Results (2007 – 2012)
Year		Trades	Win/Loss	Win/Loss %	Total Gain	Avg Gain per Trade
2012		35	30W / 5L	85.7%	$2,501	$71.45
2011		50	42W / 8L	84.0%	$4,961	$99.22
2010		42	36W / 6L	85.7%	$4,470	$106.42
2009		75	59W / 16L	78.7%	$9,099	$121.32
2008		72	54W / 18L	73.3%	$5,806	$80.64
2007		45	36W / 9L	80.0%	$4,087	$90.82
Total		319	257W / 62L	Avg 80.56%	$30,924	Avg $96.94

The lower portion of the table provides an annual assessment of the 5 strategies assuming they had been traded simultaneously since 2007. The year with the lowest Win/Loss ratio was 2008 (73.3%), the year with the highest total gain and average gain per trade was 2009 ($9,099 and $121). Also, the trades generated by the 5 strategies in both 2012 and 2010 were 85.7% winners.

So, if it sounds too good to be true, it cannot be true?

Good question. And, you are correct - it is not true..... at least, somewhat.

The three 'gotcha's' I can think of are these:

1. No one can trade GDX for free. There were 319 round trip trades which is actually 638 individual buy/sell orders. If one payed, say, $7 commission per trade, that would take $4,466 (638 X $7) off the gross gain of $30,924. 

2. Approximately 30% of the time there are 2 and even 3 strategies active. That is, sometimes the strategy signals overlapped. So, it would be entirely misleading and incorrect to suggest the cumulative sum gross gain ($30,924) was accomplished by always trading 100 shares. In fact, about one third of the time a person would have had 200 and sometimes even 300 shares in the market.

3. While the computer does tell me the trade the night before, the price it uses for calculation purposes is the opening price the following morning. If my order the following morning did not attain the opening price, that could change things some, but probably not too much.

Going forward, I plan to code these strategies into TradeStation and give them a more rigorous machine waching. If any/all of the strategies are still standing after that I will make their signals available on the website. 

I would like to apply these and many other strategies I have created to numerous ticker symbols and ultimately make their signals and their detailed statistical evidence available to everyone interested. 

I'm on a journey. Want to join me?

John
tsiTrader@gmail.com