Sunday, November 25, 2012

GDX Trading Systems - 3


Somewhat on a whim, I became curious to find out how well the 8 trading strategies discussed in the previous two posts would perform on a variety of somewhat related ticker symbols.

Mind you, these strategies were optimized to trade Market Vectors Gold Miner ETF (GDX) and never run on anything else. I was almost afraid to try this experiment because it was like flying blind - I suspected my plane would crash and burn and it would be back to that familiar and lonely blank chalkboard again.

But there were a few pleasant surprises as you will see.

What follows are the results of back testing each strategy on the entire trading history of 10 different ETFs. As the gross profit/loss would vary considerably due to the length of time an ETF had traded I devised a little metric to help me put things in perspective. 

What I did was write an indicator to measure the number of trading days each ETF has traded. Then, when I divided the gross profit/loss by this measurement I could get a 'relative' sense as to the significance of the profit/loss figure.

These are the ETFs I used along with their number of trading days tested:


ETF Ticker Symbols Tested


ETF Symbol
Description
Trading Days Since Inception
1
GDX
Market Vectors Gold Miners
1,655
2
GDXJ
Market Vectors Junior Gold Miners
   766
3
GLD
SPDR Gold Trust
2,049
4
DGP
DB Gold Double Long ETN
1,197
5
NUGT
Direxion Gold Miners BULL 3X
   496
6
DUST
Direxion Gold Miners BEAR 3X
  496
7
SLV
iShares Silver Trust
1,674
8
USLV
VelocitySharesTM 3X Long Silver ETN
   280
9
SIL
GlobalX Silver Miners
2,566
10
PALL
ETFS Physical Palladium Shares
   727

The 8 strategies will be presented in the order they were introduced in the previous two posts. For each I have added a reflective comment (basically to help me focus on how to spend my time in the future). 

Also, if the WIN % was 75%+ I highlighted it silver. If the Gross Profit divided by Trading Days was above 3.0 I highlighted it green. And if the average of the 10 ticker symbols was just less than 3.0 I highlighted it magenta. 


Strategy 1: Pivot


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$8454
100/20
83.3
5.11
2
GDXJ
$1051
45/12
78.9
1.37
3
GLD
$2,901
31/15
67.4
1.42
4
DGP
$1,548
61/34
64.2
1.29
5
NUGT
$1,024
46/20
69.7
2.06
6
DUST
$5,466
66/26
71.7
11.02
7
SLV
$1,521
61/39
61.0
0.91
8
USLV
$520
33/19
63.5
1.86
9
SIL
$3,163
135/82
62.2
1.23
10
PALL
$2,999
37/12
75.5
4.13
Average Gross Profit/Loss per Trading Day
2.85
Reflection:  Find out why GDX, DUST and PALL worked extremely well then make adjustments.


Strategy 2: GAP


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$5,380
73/15
82.9
3.25
2
GDXJ
$1,490
31/11
73.8
1.95
3
GLD
$8,794
113/29
79.6
4.29
4
DGP
$3,619
65/18
78.3
3.02
5
NUGT
$1,586
22/4
84.6
3.20
6
DUST
$2,223
26/9
74.3
4.48
7
SLV
$147
55/20
73.3
0.09
8
USLV
$1,115
11/5
68.7
3.98
9
SIL
$2,905
57/15
79.2
1.13
10
PALL
$2,018
35/11
76.1
2.78
Average Gross Profit/Loss per Trading Day
2.82
Reflection: Best overall strategy. Suspect stopLoss mechanism can be improved.


Strategy 3: PercentB


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$5,070
32/1
97.0
3.06
2
GDXJ
$416
9/2
81.8
0.54
3
GLD
$3,822
26/1
96.3
1.87
4
DGP
$921
13/1
92.9
0.77
5
NUGT
$123
2/1
66.6
0.25
6
DUST
$755
6/1
85.7
1.52
7
SLV
$1,146
18/4
81.8
0.68
8
USLV
$739
4/0
100.0
2.64
9
SIL
$179
22/5
81.5
0.07
10
PALL
$1802
12/0
100.0
2.48
Average Gross Profit/Loss per Trading Day
1.39
Reflection:  Highest Win% strategy. Suspect stopLoss mechanism can be improved considerably.


Strategy 4: Step_MA


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$6,482
38/8
82.6
3.92
2
GDXJ
$2,230
15/3
83.3
2.91
3
GLD
$4,910
47/9
83.9
2.40
4
DGP
$681
16/5
76.2
0.57
5
NUGT
$1,340
5/0
100.0
2.70
6
DUST
$793
6/2
75.0
1.60
7
SLV
$1,600
13/3
81.2
0.96
8
USLV
$1740
1/1
50.0
6.21
9
SIL
$329
17/9
65.4
0.13
10
PALL
$1327
9/5
64.3
1.83
Average Gross Profit/Loss per Trading Day
0.69
Reflection:  2nd best Win%. Needs careful study of failures and may be worth the effort.


Strategy 5: UlcerX


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$6,160
18/14
56.2
3.70
2
GDXJ
$362
4/17
19.0
0.47
3
GLD
$1,034
7/9
43.7
0.50
4
DGP
$2,038
11/16
40.7
1.70
5
NUGT
$60
4/10
28.6
0.00
6
DUST
$1,245
8/6
57.1
2.51
7
SLV
$1,637
8/9
47.1
0.98
8
USLV
$95
1/2
33.3
0.34
9
SIL
$1,258
9/28
24.3
0.49
10
PALL
$789
8/11
42.1
1.09
Average Gross Profit/Loss per Trading Day
1.11
Reflection:  Dud.


Strategy 6: Force Index


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$4,001
8/4
66.6
2.42
2
GDXJ
$596
1/1
50.0
0.78
3
GLD
$3,037
5/2
71.4
1.48
4
DGP
$1,110
2/0
100.0
0.93
5
NUGT
$173
2/4
50.0
0.35
6
DUST
$1,350
0/6
0.0
2.72
7
SLV
$80
2/3
40.0
0.05
8
USLV
$1516
0/5
0.0
5.41
9
SIL
$1077
6/12
33.3
0.42
10
PALL
$233
1/0
100.0
0.32
Average Gross Profit/Loss per Trading Day
1.38
Reflection:  Dud.



Strategy 7: VZO


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$5,606
14/5
73.7
3.39
2
GDXJ
$95
4/6
40.0
0.12
3
GLD
$4,011
15/9
62.5
1.96
4
DGP
$718
11/16
40.7
0.60
5
NUGT
$1,842
3/2
60.0
3.73
6
DUST
$597
1/6
14.3
1.20
7
SLV
$985
8/9
47.1
0.58
8
USLV
$1,666
1/2
33.3
5.95
9
SIL
$274
14/16
46.7
0.11
10
PALL
$537
3/7
30.0
0.74
Average Gross Profit/Loss per Trading Day
1.42
Reflection:  Weak. Probably not worth more effort.


Strategy 8: TSI


ETF
Symbol
Gross
Profit/Loss
 Win/Loss
Win %
Gross Profit/Loss
Divided by Trading Days
1
GDX
$6,041
18/8
69.2
3.65
2
GDXJ
$515
4/6
40.0
0.67
3
GLD
$2,026
13/13
50.0
0.99
4
DGP
$128
3/8
27.3
0.11
5
NUGT
$457
3/4
42.9
0.92
6
DUST
$1,581
2/9
18.2
3.19
7
SLV
$989
4/4
50.0
0.59
8
USLV
$311
2/1
66.6
1.10
9
SIL
$546
8/8
50.0
0.21
10
PALL
$219
4/6
40.0
0.30
Average Gross Profit/Loss per Trading Day
0.51
Reflection:  Dud.


So there you have it. Looks like 4 strategies with promise and 4 duds. What's it look like to you?

Keep smilin'

John



22 comments:

  1. Hi John
    I have been thinking about what you are doing. Its pretty amazeing the zeal you have for this and a great idea.
    I think its a worthy persuit. I just had a thought. How about a program to walk us through cycles. Enter cycle choice i.e. gold, SnP ect.
    give the cycle timeframes. then the indicators to watch for, i.e. sentement, swing, rsi,bow/sos,TSI ect.
    perhaps what the other corrisponding cycles are doing. The dollar, or, the Snp, Factor in the whining on garys blog or number of blog post (joke)
    Maybe cycles are to much of a art to do this.

    Just a thought
    jeff

    ReplyDelete
  2. Jeff - I wish your idea was possible to accomplish but I'm afraid
    it is beyond me. We are very fortunate to have Gary, as he is the true
    master of these considerations. He and I have talked some about mixing
    cycles with TSI but I have not taken it further than simple discussion.
    Maybe someday I'll figure out a way to do it. I'll keep trying.

    ReplyDelete
  3. If I may, what is the $ amount of each trade so one can see the risk/return profile.
    Thanks

    ReplyDelete
  4. Anon - hummm... that would be tough for me to figure out without either spending
    a whole lot of time or getting the numbers out of TradeStation (which I do not
    have working yet with the strategies).

    But tell you what, write me an email and tell me which strategy and ticker symbols
    you want to see the data from and I will email the Think or Swim report(s). Then
    I imagine you could run that through Excel or Google docs and have all the fun with
    it you want. Sound good? John tsiTrader@gmail.com

    ReplyDelete
  5. Sorry to say I didn't realize you were posting still. Your RSS feed doesn't have any new postings since August. Just letting you know in case it might be easily fixed. Thanks!

    ReplyDelete
  6. John, this analysis of GDX is very intriguing. I'm heavily into GDX right now and it's good to know I'm not the only one thinking miners are very undervalued. You have a wealth of knowledge here and I look forward to hearing more from you!

    ReplyDelete
  7. Thanks Mike - I heard this once before a month or so ago. Hummmm... I may have to look into it.
    First I have to figure out who I should contact. Any ideas?

    Thank you Brad. GDX has made a gigantic megaphone pattern that begins in late 2010.A couple of weeks ago
    GDX touched down on the lower boundary of this pattern then bounced. I expect that by late Winter GDX will
    reach the upper boundary of the pattern - presently at around $72. At that point gold should also be testing
    its all-time highs ($1923). As GDX is just south of $50 today, you could pocket a 40% gain in a few months
    if the bull does not succeed in throwing you off. HANG ON TIGHT! John

    ReplyDelete
  8. John, thanks for the extra thoughts on GDX. I had not noticed the megaphone formation at all.. very interesting! This bull is certainly hanging on tight.. and adding more on weakness as we move along!

    ReplyDelete
  9. It would not surprise me to see GDX get whacked one more time (unfortunately).
    Maybe a retest of that lower megaphone line - call it $46 or thereabouts.

    The reason I expect this is because the dollar has not begun its new daily cycle,
    despite giving a false start a couple days ago. Once it bottoms - any day now -
    it should go up for a short spell and that will take the miners and the stock market
    down. Heck, with a little luck you could get an entry that would materialize a 50-60%
    gain. Hummm.... I'm kinda jealous. But I am cheering for you. Go get 'em Brad!

    ReplyDelete
  10. Can you give more detail on the studies and setups (pivot, ma, gap etc). I read the earlier posts but I don't understand completely what triggers a buy or sell?

    ReplyDelete
  11. CycleTrader - I sincerely doubt, short of sending you the exact code, the answer to your question
    will be particularly helpful. But I'll let you be the judge of that.

    A horizontal range is defined to trigger both the buy and sell points when using the Pivot Point
    Moving Average. The PercentB buy and sells are also triggered by the crossing of a horizontal level,
    the upward or downward trajectory of the indicator and a consideration for whether it is accelerating
    or decelerating. An open lower gap triggers the Gap strategy to buy, while a change in the direction
    of a price moving average triggers the sell. Step uses both ATR and TSI direction for entry and exit.

    As you saw in my data, these and the other strategies worked incredibly well on the GDX lifetime, but not
    nearly as well when they encountered other ticker symbols. The challenge I have now is to figure out why
    this is the case and see if adjustments to the strategies will either make them more likely to succeed with
    GDX in the future, and even more desirable would be for them to work reliably on a whole list of ticker symbols.

    ReplyDelete
  12. John, nice call on another $46 hit. Speaking of buy/sell signals.. what about Bollinger Band crashes? GDX showed us a nice one today. Perhaps tomorrow’s action in miner’s will be good for a quick spike lower and then a kind of “inside” day. That would make it easier to create a swing low on Thursday to confirm a new daily cycle. Also, notice the last “bottom” how we had the Bollinger Band crash, and then had that inside day immediately after. We could be getting the same setup again. I added to my GDX positions this morning on the lows to DCA. http://img826.imageshack.us/img826/9931/54914016.png

    ReplyDelete
  13. It forever amazes me how Fibonacci relationships appear absolutely everywhere on gold (and GDX) charts.
    That lower megaphone boundary, I now notice, and the place where GDX has now twice recently bounced ($46)
    is also.....ta da...the 61.8% retracement of the May 16 low (aka C-wave begining).

    Gary is the Man on Bollinger Band crash trades from lots of experience and study. And I imagine he would
    agree with your analysis of the setup. I certainly do, fwiw.

    The daily TSI (7,4) on the dollar is really getting close to being way overdone - in most cases. But I have
    a sneaking suspicion the dollar may perform as it did in mid-Sept to mid-October 2010. At that time the TSI writhed
    between -50 and -70 while gold blasted higher. Also, the second IC of the previous C-wave made a spectacular
    run gaining 25% in *just 8 weeks* (late 2009). Too bad I'm out of cash because I definitely would be a buyer here,
    like yourself.

    ReplyDelete
    Replies
    1. John,

      GDX 45.55, are we there yet?

      Delete
  14. Anon - It looks like the dollar is going to take another shot at getting
    up through its 200 dma. I guess we'll need to keep the helmut on a little
    longer and count the days until it tops. Gold is now seated on the lower
    trend line of its own wedge/triangle formation - rather similar to the much
    larger specimen that it broke up through this past later August.

    Another item of trivia, the earlier and larger wedge used the July 2011 low
    to begin its lower trend line, and of course the September 2011 all-time high
    to anchor its upper trend line. Of all things, where does today's gold at $1700
    fit into this? You guessed it...we are smack at the 50% retracement. Another of
    those forever familar Fibonacci things, I suppose.

    ReplyDelete
  15. John, you lost me. First how many shares are you trading here? Second, where is the percentage return colume? ty.

    ReplyDelete
  16. Anon - sorry about that. The first article details that I am always buying and selling
    the same 100 shares. There is no % return given by Think or Swim.....just the trade dates,
    prices and a sum of gross profit or loss. When I get these strategies running on Trade Station
    *then* I will be given reams of information to share. But concerning the strategies that
    I optimized/curve fit for GDX, the trades are short enough that one is not in the market more
    than maybe 20% of the time - but that is just an eye balled guess.

    ReplyDelete
  17. John, we didn't get a big push from the BB crash just yet. But I believe next week we should see some follow through. I'm hoping gold/silver both give us a swing low on Monday. We have the potential for some big gains here, but the bulls better get busy. What do you think about an initial target of about $53 by mid Jan? Also, we may have some divergence beginning to show on the RSI and on the MACD histogram. If we can confirm a swing on Monday, I may look to enter SLV calls and add one more tranche to my GDX calls. I haven't been through the COT yet this week, but from what I have heard already, comms reduced short positions in both metals.

    Saddle up!

    http://img545.imageshack.us/img545/1695/14119341.png

    ReplyDelete
  18. Hi Brad - your GDX target and timing sound right to me.
    I think the bottom is in for gold and as soon as I can
    get some fresh money into my trading account I am going
    to load up - hopefully by the end of this week. We seemed
    to have gotten the swing low on gold on Friday and with the
    dollar now rocketing upwards it will not be long before it
    tops. Great time to buy.....now.

    ReplyDelete
  19. John.. check out the volume across the different prices of GDX Mar '13 calls from today, Mon 10 Dec. Interesting!

    ReplyDelete
  20. Brad - perhaps you would be kind enough to share what you saw
    and why you found it interesting? Thanks.

    ReplyDelete
  21. There were three major purchases of GDX calls today. The first was of 5,000 contracts at the $43 strike. The second was 10,000 contracts at the $47 strike, and the last was 5,000 contracts at the $51 strike. Now that I look at it more.. perhaps it's an options strategy that I am not familiar with. Either way, that's a lot of contracts and I've never seen volume like that on GDX calls over the last couple of years as far as I can remember.

    ReplyDelete