andy_b
Junior Member
Posts: 2
Registered: 12-6-2018
Member Is Offline
Mood: No Mood
|
|
Indicator, entry level and norm
Have taken a trial for this. So far so good. I do have a question though on one trial run:
result = ((GSB_Norm2(Momentum(Close, i1length) of Data(i1Data), 10, 100) of Data(i1Data) * i1Weight) * ((GSB_Norm2(GSB_KeltnerChannel(Close, i2length,
i2numAtrs) of Data(i1Data), 10, 100) of Data(i1Data) * i2Weight) * (GSB_Norm2(GSB_ForwardReverseEMA(i3aa) of Data(i1Data), 10, 100) of Data(i1Data) *
i3Weight)));
When I plot this in TS... the range is like -600,000 to +600,000
The decision involves:
decision = GSB_Decision(result, 1, entryParams);
where entryParams = 2000
the thing here is that 2000 seems low for the total range?
Is this normal? essentially its just looking for any time its positive + filter is true.
for reference:
norm method =10
Inputs:
i1Data(1),
i1length(38),
i2Data(1),
i2length(29),
i2numAtrs(-1.5),
i3Data(2),
i3aa(0.375),
i1Weight(-1),
i2Weight(1),
i3Weight(1),
entryParams(2000),
iSFData(3),
iSFWeight(1),
sfEntryLevel(19);
|
|
|
admin
Super Administrator
Posts: 5069
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by andy_b  | Have taken a trial for this. So far so good. I do have a question though on one trial run:
result = ((GSB_Norm2(Momentum(Close, i1length) of Data(i1Data), 10, 100) of Data(i1Data) * i1Weight) * ((GSB_Norm2(GSB_KeltnerChannel(Close, i2length,
i2numAtrs) of Data(i1Data), 10, 100) of Data(i1Data) * i2Weight) * (GSB_Norm2(GSB_ForwardReverseEMA(i3aa) of Data(i1Data), 10, 100) of Data(i1Data) *
i3Weight)));
When I plot this in TS... the range is like -600,000 to +600,000
The decision involves:
decision = GSB_Decision(result, 1, entryParams);
where entryParams = 2000
the thing here is that 2000 seems low for the total range?
Is this normal? essentially its just looking for any time its positive + filter is true.
for reference:
norm method =10
Inputs:
i1Data(1),
i1length(38),
i2Data(1),
i2length(29),
i2numAtrs(-1.5),
i3Data(2),
i3aa(0.375),
i1Weight(-1),
i2Weight(1),
i3Weight(1),
entryParams(2000),
iSFData(3),
iSFWeight(1),
sfEntryLevel(19); |
Hi Andy,
good you are going well on the trial.
The bigger the entryParams, the less trades you get, and less sensitivity to noise. I found low values worked best. You can use larger ones and see
how often they are used.
A manual optimization of the value on one system will also give you a feel for good range.
|
|
|
andy_b
Junior Member
Posts: 2
Registered: 12-6-2018
Member Is Offline
Mood: No Mood
|
|
ok then if entry params are say 0. If indicator 1, 2 3 are all over 0 and secondary filter is doing this....
Does that mean there is a really low weight placed on where all 3 indicators line up?
Essentially just looking for indicators to confirm positive, over 0... and that the secondary filter is doing this for a trade...
Lets say I was trading a RSI 2.... if its below 20 buy and sell short over 85...
Hence I was thinking where indicator 1, 2 3 line up would be important entry thresholds??
|
|
|
admin
Super Administrator
Posts: 5069
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
The weights were designed in the days where we used + instead of *
Now we use * we only need -1 or 1 for weight (to invert the signal)
All the data is normalized to average around zero. I did a lot of testing on the threshold last week.
If we use 0 or 0 to 100,000 it made no significant difference in the market validation / nth tests. Hence im ok leaving at at 0
So we are looking for them to be above or below zero to set the direction.
But then there also is entry type. Might be we look for result>0 or crossover of result or two results in a row > 0
Hope my answer helped. Nice to understand this, but not critical to get into the inner details of this part of GSB.
Im trying to make the defaults the best setting. SOme of that varies from one market to another.
|
|
|
|
![[*]](images/davis/default_icon.gif){kind=icon}
