System33 (06 Divergence signal only)

Systems home page    Open trades    Closed trades    Latest trades    Contact     Bookshop    Trivia

SYSTEM33 TEST 1 original version (Profitable 80% of times > 30 mins)

(06system33Divergence signal test)

These pages are to keep a log of systems put forward to make an automatic trading system for the futures-spread betting market.

 

1. System33 the original system without modification will retain this name, any modifications to this will carry a version number such as system33.2 for ease of understanding, and all changes will be noted on these pages.

 

SIGNALS included in this system are :-

 

For the ENTRY

 

1. 06system33divergence (using divergences of the tradevolumeindex indicator) (DEMBUY)

2. 06Insync (using the well know insync indicator with filters) For the (INSBUY)

3. support06 (JUST OVER SUPP)

4.resistance06  (JUST UNDER RESIST)

 

For the EXITS

 

5 pjlongstop

6.pjshortstop

 

Commissions for FT-SE and DAX will be set to 3 points for ALL TESTING, 1 Contract only.

 

pjshortstop_ra 3
pjshortstop_acc 0.05
pjshortstop_artlen 8
pjlongstop=ra 3
pjlongstop_acc 0.05
pjlongstop_artlen 8
SYSTEM33divergence_trigger 1
Commissions 3
   

 

FUNCTIONS Used

 

This system decides on the overall direction of trend by using 3 indicators which go to construct the function called system33

Preroctrendlt is based on momentum and asks if it is <>0 to allocate trend scores, if > 0 then it assigns a value of 0.1 * 100 day momentum.

Nearhilo has 3 different lengths

Schwartz trendperc again uses 3 lengths to determine trend

The trend score 1 or -1 is uses to define the direction of long term trend, and this needs to be tested on different settings.

 

INDICATORS

 

This can view simply by using the indicator called SYSTEM33

 

DATA Used

 

H.S FT-SE100 Futures (with many gaps) so this test is for preliminary testing only

 

DATA-type TICKCOUNT!

 

First date November 13th 2001

Last date August 17th 2007

 

TESTING WITH EACH SIGNAL PRODUCED THE FOLLOWING RESULTS

 

 

TEST 1

 

 06system33divergence (using divergences of the tradevolumeindex indicator) (DEMBUY)

 

TIMEFRAMES (minute) note, if maxdd was caused by bad data, then # will be used.

Equity curve will be rated like an exam, if drawdowns more than 700 = fail.

 

All P-L and DD factors defined in POINTS.

 

N. Interval Trades Profit Win-lose  Risk-R  Max DD  Ave win  Ave loss  Best win Worst loss Kelly Equitycurve
  Week 0                    
C+ Daily 6 409 66.6% 1.03 -491 198 -192 484 -344 34 DD large
C 600 20 242 55% 1.00 -783 123 -315 267 -315 10 DD FAIL
  550 18 554 50% 1.88 -272 131 -69 530 -201 23 DD not bad
B+ 500 52 1178 52% 1.40 -582 129 -92 372 -257 17 DD Large
  450 15 -70 46% 1.08 -765 171 -158 447 -334 NEG  LOSING FAIL
B- 350 14 730 64% 1.24 -287 147 118 459 -255 34 DD Marginal
B+ 240 23 1105 60.8% 1.39 -599 1147 -106 352 -259 32 DD large
  180 43 675 53.49% 1.20 -680 107 -89 275 -181 13 DD large
B+ 150 41 1545 56.1% 1.49 -738 141 -95 441 -267 26 DD Fail but ok
  120 64 310 45.31% 1.35 -1004 101 -75 313 -253 4.7 DD Fail
  100 97 -12 45.3% 1.20 -1403 85 -220 296 -220 NEG DD Fail LOSS
C 80 103 881 47.5% 1.41 -852 83 -59 246 -187 10 C > 2003
D 60 135 513 47% 1.25 -892 71 -56 220 -197 0.5 D > 2004 DDFAIL
C 50 171 345 43.2% 1.39 -1188 78 -171 650 -171 1.9 DD Fail but ok
  40 194 -66 42.7% 1.32 -1659 60 -145 294 -166 NEG DD Fail
B- 30 279 1151 41.22% 1.67 -1293 68 41 388 -149 5.9 DD Fail but ok
  20 429 -94 41.49 1.39 -1013 46 -33 244 -128 NEG NEG < 2003
  15  608 -870  41%  1.31 -1418 38 -29 229 -175 NEG NEG
  10 slow   loading   chart            
  5 slow   loading   chart            

 

Comments for anomyly at 500 mins why so many trades? perhaps it found the right frequency?

 

 

Comments for times daily to 120 mins.

One initial testing, it appears stats can be improved by using a tighter stop and possibly incorporating a max-loss stop as protection, as this system has many fine signals at or close to the bottom of the move.

 

Comments for  80 mins.

Produced a nice curve after 2003, but lost money > 500 points in 2007 dump. Worth checking stops.

 

Comments for 60 mins.

Produced nice curve after march 2004 worst dd -200 there after. Highlighted cells show reduced testing data set from October 2004 very nice results.

 

Comments for times  50 mins. Lost 800 points 2001-2004 then made 1400 points!

 

TEST 1 Improvement ideas.

 

Optimise stop in the most promising time frames 700 minutes down to 30 minutes 10 min increments.

Add a max loss stop;

Add a "close trades at end of day"

Add a " dont trade later than 3pm"

This will done in page System33_optimise