Forums  > Risk Management  > deposit protection - model unstability / programming error  
     
Page 1 of 1
Display using:  

svisstack


Total Posts: 303
Joined: Feb 2014
 
Posted: 2015-08-19 23:10
[please move this to trashcan, it dont contain any valuable information's]

Time well wasted.

goldorak


Total Posts: 1000
Joined: Nov 2004
 
Posted: 2015-08-20 06:39
I would use an external, totally different software. I guess here relying on algos that typically do networks monitoring is a good start. Typically the algo would learn from your trading systems' behavior and be able to give you an alarm and, if you have big balls, solve the problem by itself (or herself depending on how much you are in love with your algo) by taking independent actions.

Not an easy task though.

If you are not living on the edge you are taking up too much space.

katastrofa


Total Posts: 362
Joined: Jul 2008
 
Posted: 2015-08-20 11:53
"few months ago I have case that my strategy who was working few months already in real-time environment in some very special cases, caused by programming error went haywire and placing aggressive orders around spread on various side of it"

Strategy orders should be checked by a separate piece of software which would reject obviously "bad" orders (e.g. bid crossing the ask, etc.)

svisstack


Total Posts: 303
Joined: Feb 2014
 
Posted: 2015-08-20 14:13
.

Time well wasted.

TSWP


Total Posts: 382
Joined: May 2012
 
Posted: 2015-08-20 17:15
svisstack, in theory all possible known issues can be categorized. If you can categorize issues you should be able to implement a solution. Lot of work probably if you have many issues.
If an issue cannot be categorized it means it is (yet) unknown and you cannot do anything about it until it becomes known (in a way a terrifying scrum approach, possibly fatal in a field like finance but that is the cost of complexity).

If I had to choose a path: simplify as much as possible each detail of your strategy and related code because the higher the degree of complexity and the harder it becomes to control everything and the more the risk of errors, I am stating the obvious sorry.

"Acceptance by the machine required absolute precision on my part. A mistress of perfect consistency, the computer rejects all but the flawless, offering no explanation.When the acceptable is finally offered, the machine's acceptance is total, unwavering and eternal."
[Paul Lutus - NASA and "Apple Writer" programmer, not a trader]




"I couldn’t write programs […] I wouldn’t have done well in an Olympiad or a math contest. But I like to ponder. And pondering things, just sort of thinking about it and thinking about it, turns out to be a pretty good approach." - Jim Simons
Previous Thread :: Next Thread 
Page 1 of 1