JasonT
Junior Member
Posts: 61
Registered: 6-6-2018
Member Is Offline
Mood: No Mood
|
|
Exits
Hi Peter,
when using exit types such as a Stop Loss (eg $1250) or Profit Target (eg $625) how does GSB determine (estimate) the outcome of a trade, especially
on a 30 minute or larger bar which can have large ranges at times? I'm assuming that GSB does not use decades worth of intra-bar tick data to
recreate what would have actually happened, in what sequence per bar and so some form of estimation is expected.
Ultimately I think I want to understand what should be considered when using these types of exits to build systems with GSB.
Many thanks.
Jason
P.S. I ran GSB using a $250 stop loss and Market On Day Close. I exported the trades from one strategy that was created. I noticed that some losing
trades lost more than the $250 stop loss (see attached image). Is this supposed to happen like that? I guess this is relevant to the broader
question above. Thanks.
|
|
|
admin
Super Administrator
Posts: 5069
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by JasonT  | Hi Peter,
when using exit types such as a Stop Loss (eg $1250) or Profit Target (eg $625) how does GSB determine (estimate) the outcome of a trade, especially
on a 30 minute or larger bar which can have large ranges at times? I'm assuming that GSB does not use decades worth of intra-bar tick data to
recreate what would have actually happened, in what sequence per bar and so some form of estimation is expected.
Ultimately I think I want to understand what should be considered when using these types of exits to build systems with GSB.
Many thanks.
Jason
P.S. I ran GSB using a $250 stop loss and Market On Day Close. I exported the trades from one strategy that was created. I noticed that some losing
trades lost more than the $250 stop loss (see attached image). Is this supposed to happen like that? I guess this is relevant to the broader
question above. Thanks.
|
GSB & TS will use the open price if there is a gap which can be worse than the actual stop price. This is a fair reflection of what happens in live
trading. Im not a fan of using such tight stops and GSB may just re-enter after being stopped any one. There is some evidence that this is actually a
good thing. Personally I build systems with no stops, and put one on when done. This is just my preference and lots of users dont do this. I think
your stops and PT are way too low for ES.
|
|
|
JasonT
Junior Member
Posts: 61
Registered: 6-6-2018
Member Is Offline
Mood: No Mood
|
|
Hi Peter,
thanks for the explanation of the method. Yes I agree that $250 is too tight a stop for ES.
I was running a test to see how GSB handled different stop loss amounts (eg $2000, $1000, $500, $250 etc) and their effect on system performance. In
general, as the stop loss got smaller the best equity curves got straighter and smoother. From my experience in testing with Ninjatrader this is a
warning sign, as typically tighter stops produce worse performance, which I think is the point you are making.
My personal view is that autotrading with no stop loss in place exposes all of your capital and doesn't protect against large price moves that can
happen intraday on some markets. For me its not acceptable to leave such a risk unmanaged. That's the reason for testing the fixed stop amounts.
If a trader's view is that they must use a stop loss, whatever value is selected, then its affect on a system they intend to trade needs to be tested,
because it will affect a system's performance. It might improve the performance but it also might render it unprofitable.
Is GSB able to measure Maximum Adverse Excursion?
That might be a way to identify how much risk the system would have been exposed to and how often.
Thanks 
|
|
|
admin
Super Administrator
Posts: 5069
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by JasonT | Hi Peter,
thanks for the explanation of the method. Yes I agree that $250 is too tight a stop for ES.
I was running a test to see how GSB handled different stop loss amounts (eg $2000, $1000, $500, $250 etc) and their effect on system performance. In
general, as the stop loss got smaller the best equity curves got straighter and smoother. From my experience in testing with Ninjatrader this is a
warning sign, as typically tighter stops produce worse performance, which I think is the point you are making.
My personal view is that autotrading with no stop loss in place exposes all of your capital and doesn't protect against large price moves that can
happen intraday on some markets. For me its not acceptable to leave such a risk unmanaged. That's the reason for testing the fixed stop amounts.
If a trader's view is that they must use a stop loss, whatever value is selected, then its affect on a system they intend to trade needs to be tested,
because it will affect a system's performance. It might improve the performance but it also might render it unprofitable.
Is GSB able to measure Maximum Adverse Excursion?
That might be a way to identify how much risk the system would have been exposed to and how often.
Thanks
|
TS can do that. Stops are normally not critical as long as they are not too tight.
|
|
|
kelsotrader
Junior Member
Posts: 29
Registered: 16-2-2018
Location: Tapanui - New Zealand
Member Is Offline
Mood: No Mood
|
|
I have just found another issue with having a tight stop loss.
The current default for max trades per day is blank ie as many as meets the entry parameters.
That means that the system can keep entering trades and getting stopped out.
I know because I have seen it happen.
The answer is to make sure that trades per day are limited.
I would suggest that everyone enter 1 into Maximum Trades Per Day to ensure it is limited and not just taken for granted.
|
|
|
admin
Super Administrator
Posts: 5069
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by kelsotrader | I have just found another issue with having a tight stop loss.
The current default for max trades per day is blank ie as many as meets the entry parameters.
That means that the system can keep entering trades and getting stopped out.
I know because I have seen it happen.
The answer is to make sure that trades per day are limited.
I would suggest that everyone enter 1 into Maximum Trades Per Day to ensure it is limited and not just taken for granted. |
This is an interesting topic. I coded what you said in my systems some time ago.
However there is merit to allow say 3 trades per day with a much tighter stop.
if the signal to enter disappears as the market is going against you, your losses are reduced.
There is however some sort of bug (intermittent??) in gsb stop code that we hope to fix very soon. Other exits are also being worked on now too, but
priority is bug fixes.
|
|
|
kelsotrader
Junior Member
Posts: 29
Registered: 16-2-2018
Location: Tapanui - New Zealand
Member Is Offline
Mood: No Mood
|
|
I noticed this issue for the first time last night with ES . The strategy had a $350 stop loss. It did two losing trades and I intervened. Had I not
intervened and a third trade in the same direction had been started it would have broken even or a slight profit.
I was unaware that more than 1 trades could be traded in one day. My fault for not paying enough attention to settings in GSB and assuming only one
trade per day was traded.
I agree that some limit on the number of trades is essential when using a stop loss.
|
|
|
![[*]](images/davis/default_icon.gif){kind=icon}
