<- Back to Tasks

Calculating SQN

Můžete prosím ověřit výpočet hodnoty SQN v SQX ? Ve starém SQ to bylo v pořádku. V SQX mi vychází, že se s rostoucím počtem obchodů SQN snižuje a mělo by to být opačně.

Děkuji

Attachments
No attachments
  • Votes +2
  • Project StrategyQuant X
  • Type Bug
  • Status Refused
  • Priority Normal

History

PK
#1

investor666

04.06.2019 16:56

Task created

h
#2

hankeys

05.06.2019 09:52
Voted for this task.
PK
#3

investor666

05.06.2019 10:23
Voted for this task.
MF
#4

Mark Fric

07.06.2019 09:01

Status changed from New to Waiting for information

please write in English also in description, next time it will be rejected.


I checked SQN computation, but I haven't found any error. Could you attach few strategies where you think it is suspicious?

PK
#5

investor666

07.06.2019 09:23

Attachment Capture SQN1.gif added

For me it is very unusual, that strategies with the same Sharpe ratio and with more trades have smaller SQN that the others. 

I think that calculation Sharpe ratio and SQN is very similar - SQN  is only multiplied by the square root of the number of trades ...

  

PK
#6

investor666

07.06.2019 09:30

Attachment Capture SQN3.gif added

From old SQ3
MF
#7

Mark Fric

07.06.2019 09:49
the difference between Sharpe and SQN computation in SQ X is that:


Sharpe Ratio is computed from daily results (P/L per day) - so it is not that dependent on number of trades, but on number of days

SQN is computed from individual trades, not daily results


Sharpe Ratio in SQ3 was using older version of algorithm which was fixed in SQ X.

PK
#8

investor666

07.06.2019 10:23
OK, I accept some changes of algorithm. But look at all the table from SQX - it is nonsense that commonly - strategies with more trades have smaller SQN ! Commonly ...

The SQN formula includes multiplying by the square root of the number of trades. It is a principle.   

Thanks for taking into account. 



MF
#9

Mark Fric

07.06.2019 12:13
as far as I remember, standard SQN formula as defined by Van Tharp doesn't includes multiplying by square root of number of trades.


There is SQN Score that does this, so it is independent on number of trades.

PK
#10

investor666

07.06.2019 13:53

Attachment Capture SQN4.gif added

Look please at this SQN a SQN score from SQX. Is it possible ???

Blue row = SQN score 1,07 and SQN 1,17. 

Against that (under) SQN score 0,69 and SQN 1,76.  

Could you give me please your version of these formulas (SQN and SQN Score) ? 


MF
#11

Mark Fric

07.06.2019 14:01

why shouldn't it be possible?


formulas for all stats values can be seen in Code editor in Snippets -> SQ -> Columns -> Databanks
MF
#12

Mark Fric

10.06.2019 07:14

Status changed from Waiting for information to Refused

Votes: +2

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please