Simulation Bugs

Bugs

Priority Key: 10 = Critical, 1 = Slightly Important

 

Bug ref no. Priority Description
1 8 Something is wrong as the 0.3 setting, which would normally be unusable is making £52 billion quid...

The faster models are tripping up somewhere causing huge profits to be recorded which should be losses.

We need to run a 1% stop distance and see where the error is occuring.

2 10

Big bug when running 50% system, symbol CCL computed a stop of 504974.3 from an entry at 861 and a bid of 1347.64 and as we are now

using MAX to calculate the profit that gets booked, it booked £665,429 profit ( incorrectly on that trade ) date 01/07/2002

This bug seems unique to wider stop models.

(see image bug_ccl.jpg) on quadcore pc

3 5

Day counter was at 75 or 76 then went down to around 5, then back up to 77 again, (same data bed as before around 16th 04-2000)

later it jumped by several hundred points without explanation. for many dates after it is jumping all over the place, and so is the days tested window.

(see video bug_day_counter) on quadcore pc

4 10 The system is putting on all longs at the start ( which is probably what i said we should do initially in 2005) this time i want it to wait for a stance change before doing anything. 
5 7

 

  1. BUG TO NOTE. RAN DATA COLLECTION BUT NOT INDICATOR BUILDER , OPENED SIMULATION TO TEST DONCHIAN, BUT IT SAID NO DATA TO SIMULATE.

 

2nd time loaded up some 0.4 data and the donchian crashed after 2720 days with mess no row at position 0.