Sunday, August 26, 2012

Real-Time TSI Trading System Test


Over the past few weeks I have developed a True Strength Index (TSI) related indicator/strategy that shows great promise when run through numerous back test parameters. At this point I would like to find out how it behaves in real-time. In the quest to satisfy my curiosity, I intend to update the table below for new signals generated by the system after each day's trade has closed.

At the end of this 3-week experiment I should have a sense for whether further real-time testing is needed to draw a conclusion about the system, or whether it is either a pure dud or an inspired strategy (unlikely, but I can hope)

And if you choose to follow the daily updates I am quite sure you have your own opinion of the system, which I invite you to share.



Here are 56 Futures Contracts available to me using the ThinkorSwim trading platform.

I have created a new indicator/strategy that back tests well but I do not know how well it tests in real-time.
My intention is to update this table after each day’s close for a period of 3 weeks. I expect that this real-time
documentation will help me determine how well the system works.

THIS IS AN EXPERIMENT – DO NOT TRADE THESE SIGNALS AS I HAVE NO IDEA HOW RELIABLE THEY WILL BE!

I will only enter new data within a cell if there is a new signal. Cells that are blank mean the existing signal remains
in effect. Red = Sell, Green = Buy

At the conclusion, assuming I can keep up with this for 3 weeks, I intend to summarize the findings of this EXPERIMENT - for both the good
and the bad outcomes, and offer an opinion about the system’s reliability and profitability.

John
tsiTrader@gmail.com                                                                                                                           Most recent update:  Saturday September 1



FUTURES
CONTRACT
TICKER
SYMBOL
BEGIN

WEEK 1

WEEK 2

WEEK 3
Aug 24

Aug 27
Aug 28
Aug 29
Aug 30
Aug 31

Sep 3
Sep 4
Sep 5
Sep 6
Sep 7

Sep 10
Sep 11
Sep 12
Sep 13
Sep 14
1.
Australian Dollar
6A
8/13
1.0487

1.0351
1.0358
1.0332
1.0277
1.0314












2.
British Pound
6B
8/24
1.5802

1.5796
1.5814
1.5832
1.5787
1.5877












3.
Canadian Dollar
6C
8/22
1.0083

1.0087
1.0114
1.0105
1.0071
8/31
1.0144












4.
Euro FX
6E
7/25
1.2166

1.2502
1.2569
1.2533
8/30
1.2508
8/31
1.2585












5.
Japanese Yen
6J
8/1
0.012758

8/22
0.012733
0.012741
0.012711
0.01272
0.012777












6.
Mexican Peso
6M
8/13
0.075725

0.075675
0.0757
0.075025
0.07485
8/31
0.075675












7.
New Zealand Dollar
6N
8/7
0.8133

0.8073
0.8036
0.7997
0.7974
8/31
0.8025












8.
Swiss Franc
6S
7/25
1.0135

1.0409
1.0471
1.0437
8/30
1.042
8/31
1.048












9.
Cocoa
CC
8/13
2388

8/27
2472
2584
2580
2613
2611












10.
Light Sweet Crude Oil
CL
8/23
96.04

95.69
96.14
95.18
94.77
8/31
96.56












11.
Cotton
CT
8/10
73.02

76.04
75.85
76.53
76.94
77.26












12.
U.S. Dollar Index
DX
7/13
83.44

81.71
81.36
81.58
8/30
81.725
8/31
81.195












13.
E-Mini Euro FX
E7
7/25
1.2158

1.252
1.2568
1.2534
8/30
1.251
8/31
1.2584












14.
E-Mini S&P MidCap 400
EMD
8/23
966.7

968.9
972.7
971.9
965.3
971.5












15.
E-Mini S&P 500
ES
8/23
1400

1408
1407.75
1407.5
1396.25
1404.75












16.
Gold
GC
7/25
1603.5

1666
1669
8/29
1658.7
1657.6
8/31
1692.6












17.
Eurodollar
GE
8/22
99.5925

99.6
99.605
99.605
99.6025
99.6125












18.
Feeder Cattle
GF
7/18
137.2

8/27
143.95
8/28
144.525
145.0
145.2
146.825












19.
Libor one-month
GLB
8/6
99.77

99.7675
99.7675
99.7675
99.7675
8/31
99.775












20.
Lean Hog
HE
8/13
77.1

73.1
73.125
73.85
74.225
74.1












21.
Copper
HG
8/3
3.365

3.4675
3.461
3.438
3.442
3.4615












22.
Heating Oil
HO
8/3
2.9289

3.1303
3.1257
3.1295
3.14
3.1815












23.
E-Mini Japanese Yen
J7
8/22
0.012728

0.012703
0.01274
0.012721
0.012725
0.012778












24.
Coffee
KC
8/24
163.1

166.75
167.9
165.95
163.7
164.35












25.
Random Length Lumber
LBS
8/16
309.5

290.8
286.5
286.9
288
289












26.
Live Cattle
LE
8/16
125.425

123.35
123.825
8/29
125.625
125.8
126.125












27.
E-Micro AUD/USD
M6A
8/13
1.0488

1.0351
1.0359
1.0334
1.0279
1.0314












28.
E-Micro GPB/USD
M6B
6/4
1.538

8/27
1.5794
8/28
1.5818
1.5827
8/30
1.5782
8/31
1.5879












29.
E-Micro EUR/USD
M6E
7/25
1.2167

1.2501
1.2568
1.2529
8/30
1.2509
8/31
1.2583












30.
Natural Gas
NG
8/21
2.83

8/27
2.664
2.608
8/29
2.691
2.76
2.802












31.
Nikkei 225 Stock Average Index
NKD
8/22
9085

9100
9100
9065
8905
8895












32.
E-Mini Nasdaq 100 Index
NQ
8/23
2762.75

2783
2783
2781.25
2754.25
2770.5












33.
Frozen Concentrate Orange Juice
OJ
8/24
119

113.7
113.7
117.85
117.95
118.65












34.
Palladium
PA
6/29
582.2

649.5
649.5
8/29
631.4
620.25
629.9












35.
Platinum
PL
6/29
1452.1

1543
1519.7
8/29
1517
1508
1540.9












36.
miNy Natural Gas
QG
8/21
2.83

8/27
2.665
2.608
8/29
2.695
2.76
2.80












37.
NYMEX miNY Light Sweet Crude Oil
QM
8/23
96.05

95.725
96.125
95.25
94.825
8/31
96.575












38.
RBOB Gasoline
RB
8/24
2.9069

2.9553
2.9359
2.922
2.915
8/31
2.9715












39.
Sugar
SB
7/24
23.48

19.59
8/28
20.16
19.69
19.81
19.78












40.
Silver
SI
5/17
28.015

30.67
30.85
30.775
8/30
30.49
8/31
31.79












41.
Russell 2000 Index  Mini-size
TF
8/23
805.2

809.6
813.8
814.1
805.9
811.1












42.
Ultra T-Bond
UB
8/22
166’21

166’18
167
166’17
167’06
169’17












43.
Mini-sized Gold
YG
6/29
1598.2

1666
1669.3
8/29
1658.7
1658.1
8/31
1694.1












44.
Mini-sized Silver
YI
6/29
27.43

30.67
30.821
30.725
8/30
30.461
8/31
31.767












45.
Mini Dow Jones Industrial Average
YM
8/21
13199

13104
13086
13086
12985
13082












46.
30-Year U.S. Treasury Bond
ZB
8/22
148

149’23
150
149’26
150’12
151’21












47.
Corn
ZC
7/23
785.5

800.25
796.25
8/29
814
808
8/31
800












48.
5-Year U.S. Treasury Note
ZF
8/22
124’130

8/27
124’030
124’045
124’050
124’100
8/31
124’225












49.
Soybean Oil
ZL
8/9
53.17

56.6
56.4
57.38
57.25
56.96












50.
Soybean Meal
ZM
8/15
493.7

519.6
523
529.6
534.7
533












51.
10-Year U.S. Treasury Note
ZN
8/22
133’165

8/27
132’220
132’270
132’250
133’025
8/31
133’260












52.
Oats
ZO
8/16
383.75

387
382.25
396.75
395.5
396.5












53.
Thirty-Day Fed Funds
ZQ
8/14
99.86

99.87
99.865
8/29
99.865
99.865
99.865












54.
Soybean
ZS
8/15
1634.5

1720.75
1724.25
1752
1761.25
1754












55.
2-Year U.S. Treasury Note
ZT
8/22
110’065

110’067
110’067
8/29
110’06
8/30
110’072
110’095












56.
Wheat
ZW
8/20
879.75

881.5
876
906.5
903
888.75





































11 comments:

  1. This is exciting, John. But I wonder, why futures instead of etf's? Do you suspect the outcome (whatever it is) will be different for futures and etf's?

    Joe

    ReplyDelete
  2. Hi Joe - whatever the outcome I believe the system would test the same with ETFs.
    It was easier for me to go down a pre-made and diversified list of futures than monkey
    around making a list of ETFs. Basically, this system either works or it doesn't. It's
    about that simple. If I already knew the answer I would not waste my or anyone else's
    time with this experiment.

    ReplyDelete
  3. John, very ambiguous project (need help?)! But could be very exciting! One favor (if not to much trouble), could you add a column for last daily price? Are the signals generated on daily or intraday data?
    Looking forward to seeing the daily results.

    Gerald

    ReplyDelete
  4. Hi Gerald - I suppose as I am checking each of the 56 mouse traps after each day's close
    I could add that day's closing price (black) to the empty white box. If there is a *new* signal
    that day it would, as planned already, be either red or green. The signals can be generated
    on any time frame. The bigger issue will be what to make of their statistical reliability.
    If there is not a profitable strategy to handle this issue then it's 'back to the drawing
    board'.

    ReplyDelete
  5. John,

    I applaud your ambition and laying it out for open viewing. I have two questions at this time. Is the signal generated during the day or is it only on the close? Since futures markets close at different times are you taking each at their close or just grabbing them all at the same time.

    I have traded wheat off and on for a few years and am always interested in new tools to help the cause. :)

    Thanks again for posting this.

    Chi-Town Deadhead

    ReplyDelete
  6. Chi-Town - I appreciate your polite applause and hope I can come up with something that
    works. At the moment, with so many reversal signals on Friday, I'm not so sure I have.
    But I do need to give it time.

    Yes, I try to very careful to get the day's closing price correct for each futures contract.
    I usually find a single error for the previous day somewhere when I enter the next day's closing
    numbers. I guess works as something like a double check from my side, and certainly anyone who
    spots an error - no matter how small - is encouraged to graciously point it out to me so that I
    can correct it.

    I use the signal that is present at the day's closing bell.

    The /zw daily 5 yrs Short strategy back test was +$83.4k and Long strategy was +$86.3k
    This was using a single (1) contract for each trade. I guess that's pretty good. (?)

    ReplyDelete
  7. any chance you could put in the SPY? I have had some good luck trading 3X shares against the SPY using TSI based indicators, but not with futures or after-hour data (too light of volume).

    much thanks!

    ReplyDelete
  8. John,

    That is excellent.

    A very simple strategy I used when I started was good for about 15 to 20k a year with one contract. I have had success with a number of strategies until they didn't work. :) It is a continuous battle because the grains are a very fickle group. I have a number of friends who are still floor traders in different commodities but all the grain traders have moved on to different careers. This helps a little as they don't mind sharing some insights as to how they used to do it and the pitfalls. The guys who are still trading are very closed lipped about any advantage they have. I don't blame them as it is their livelihood on the line too.

    One of the things I can tell you, at least in wheat is to watch which contract you are checking the last two weeks of the month before the contract expiration. An example is the September contract. I would already be moving over to the December contract by August 15 at the latest. Volumes drop as the big guys move to the next monthly contract. Also, whichever broker anyone has chosen will choose different deadlines to buy/sell your contract for that month automatically. Some of them want you out by the 15th, others will let you go until the final Friday of the month prior to the contract month. Most do not let you take delivery. (Which is a good thing since where are you storing 5000 bushels of wheat?) :0 So if you are checking the closing price of September wheat right now, your testing will be incorrect as almost no one would be able to get this price unless you are taking delivery or actually selling the wheat. You need to be checking the December price.

    This is where doing a large data dump for all commodities can become very involved and confusing. If, after this 3 week test you find it still favorable, maybe have some of your regulars test a certain area a little more in depth for you. Someone could do forex, another could do metals, another treasuries, you get the idea. That way if some of the regulars have an area of interest or experience they could help. I would offer to do it for grains because I don't know much about many of the others. Just a thought.

    I hope your ongoing results continue to show great promise as I will be following it. If you have any other questions I will try to answer them as much as possible.

    Thanks again for the blog,

    Chi-Town Deadhead

    ReplyDelete
  9. kcountry - unfortunately, I have more ideas than I have time to explore. I think the /ES
    contract should equate well to your interest in SPY. Please, do not make any trades based
    on what you see on this spreadsheet. I will take the data that is accumulated after three
    weeks - assuming I continue this that long - and I will try to figure out where the system
    fails and whether there is anything I can do about it. I certainly am not trading this
    information, you should be sure of that.

    Chi-Town - thank you for the interesting fine points that I certainly was not aware of. fwiw,
    thinkorswim is showing me the December /ZW (wheat) contract so perhaps it also knows how the
    game works. Yes, having assistance is always a luxury that I appreciate. Thank you for reminding
    me of that possibility. Well, I'm home from work now and have to get cracking on this before dinner.

    ReplyDelete
  10. kcounty,

    S&P E-minis (ES) are one of the most heavily traded futures so there is, so volume is not the issue. The Forex (currency pairs) is larger but talk about being chum in the water with every shark waiting to attack you. It's the margin that is the issue with futures. If you have never traded futures, the ES is where I would start, but with a practice account. It is fun and can be highly addictive. From personal experience, no matter how much you practice it will never be the same as the real thing when your money is on the line. If you have any questions please ask.

    Chi-Town Deadhead

    ReplyDelete
  11. John,

    I'm sure it has jumped over as TOS will not let you trade the September contract anymore as they will not let you stand for delivery.

    All I was saying is that if this works for say 5 of the symbols to make sure you have the nuances of that particular product when you go to the next testing phase.

    Good luck,

    Chi-Town

    ReplyDelete