log in     sign up for free
 
The more disciplined you can get, the better you are going to do in the market. The more you listen to tips and rumors, the more money you're likely to lose.
David Ryan
 
 
Home
systems (284)
 
Special offer: buy MetaStock (try for free)   |   Reuters QuoteCenter Real-Time Data (get a free month)   |   EOD Data
 
Formula Library
 
AmiBroker
MetaTrader
MetaStock
TradeStation
Wealth-Lab
NinjaTrader
 
add your formula
System Library
 
 
Account login
Your email:
Password:
 
Log in
New to s4t? Sign up.
If you forgot your password click here.
Search the Web
 
Custom Search

NLR Indicator

Formula for: TradeStation

indicator


 

 

Views:  2007

Added: June 24, 2008
 
Rate this code:  Rate: 1 Rate: 2 Rate: 3 Rate: 4 Rate: 5
This formula has not been rated yet
 

email this link

 
 
Tags: TradeStation, indicator
 
Your Ad Here
The NLR Indicator uses non-linear regression analysis to estimate a curve through price data. Potential applications include but are not limited to:
1) estimates of next price(s)
2) price envelopes and/or channels
3) moving average replacements



Type : Function, Name : NLR

{ NLR Function WAV 11/28/03

NonLinear Regression Analysis
y = A + B*X + C*X*X (2nd order)

y price
x bar (time)
A,B,C coefficients to determine

Standard deviation (S) of price from
calculated price for length data is returned

price for desired bar (either a past or future one)
is returned

}
InPuts: Length(NumericSimple), price(NumericSeries),
DesiredBar(NumericSimple),oNLRPrice(NumericRef),oStdDev(NumericRef);

vars: AvgX(0),AvgY(0),N(0),XM(0),YM(0),XM2(0),SXX(0),SXY(0),SYY(0),
SSYY(0),SXX2(0),SX2X2(0),SYX2(0),ACoeff(0),BCoeff(0),CCoeff(0),
X(0),Y(0), MaxLength(100),s(0),TX(0),TY(0),RV(0),ERR(0);

array: YValue[100](0),XValue[100](0);

if BarNumber = 1 then
begin
N = length;
{ be sure array size is protected }
if N > MaxLength then
N = MaxLength;
if N < 2 then
N = 2;
end;

{ fill arrays }
for value1 = 1 to N
begin
XValue[value1] = value1-1;
YValue[value1] = Price[value1-1];
end;

{ calc averages of x,y pairs }
AvgX = 0; AvgY = 0;
for value1 = 1 to N
begin
AvgX = AvgX + XValue[value1];
AvgY = AvgY + YValue[value1];
end;
if N <> 0 then
begin
AvgX = AvgX / N;
AvgY = AvgY / N;
end;

{ do regression and calc 3 coefficients }
SXX = 0; SXY = 0;SYY = 0;SXX2 = 0;SX2X2 = 0;SYX2 = 0;
for value1 = 1 to N
begin
XM = XValue[value1] - AvgX;
YM = YValue[value1] - AvgY;
XM2 = XValue[value1]*XValue[value1] - AvgX*AvgX;
SXX = SXX + XM*XM;
SXY = SXY + XM*YM;
SYY = SYY + YM*YM;
SXX2 = SXX2 + XM*XM2;
SX2X2 = SX2X2 + XM2*XM2;
SYX2 = SYX2 + YM*XM2;
end;

value99 = SXX*SX2X2 - SXX2*SXX2;
if value99 <> 0 then
begin
BCoeff = ( SXY*SX2X2 - SYX2*SXX2 ) / value99 ;
CCoeff = ( SXX*SYX2 - SXX2*SXY ) / value99;
end;

ACoeff = AvgY - BCoeff*AvgX - CCoeff*AvgX*AvgX;

{calc estimated price for desired bar
DesiredBar: use + for past and - for future
eg., for next bar in future:
DesiredBar = -1 }
oNLRPrice = ACoeff + BCoeff*DesiredBar+
CCoeff*DesiredBar*DesiredBar;

{calc std dev}
S = 0; AvgY = 0;
for value1 = 1 to N
begin
TY = YValue[value1];
TX = XValue[value1];
RV = ACoeff + BCoeff*XValue[value1]+
CCoeff*XValue[value1]*XValue[value1];
ERR = TY - RV;
AvgY = AvgY + TY;
S = S + ERR*ERR;
end;
if N-1 <> 0 then
if S /(N-1) > 0 then
S = SquareRoot(S /(N-1));

oStdDev = s;

NLR = 1;




The standard deviation of price versus calculated price (from NLR curve data) is easily determined. Interestingly, when the price calculation is close to the actual for several bars in a row (ie., low std dev) a significant price move sometimes occurs soon after demonstrating a possible new volatility measurement. See pic below for paintbar examples.

Code has been updated to use the new NLR function.



Type : Indicator, Name : NLR1 Indicator

{ NLR1 Indicator WAV 11/26/03

NonLinear Regression Analysis
y = A + B*X + C*X*X (2nd order)

y price
x bar (time)
A,B,C coefficients to determine

modified 11/29
utilize NLR function
added price input
removed mult input
plots only last bar from curve

}
InPuts:
Length(20), price(close);

vars: S(0),oNLRPrice(0);

value1 = NLR(length, price, 0, oNLRPrice, S);

plot1(oNLRPrice, "NLR");



 

 



Code:

Type : Function, Name : NLR

{ NLR Function WAV 11/28/03

NonLinear Regression Analysis
y = A + B*X + C*X*X (2nd order)

y price
x bar (time)
A,B,C coefficients to determine

Standard deviation (S) of price from
calculated price for length data is returned

price for desired bar (either a past or future one)
is returned

}
InPuts: Length(NumericSimple), price(NumericSeries),
DesiredBar(NumericSimple),oNLRPrice(NumericRef),oStdDev(NumericRef);

vars: AvgX(0),AvgY(0),N(0),XM(0),YM(0),XM2(0),SXX(0),SXY(0),SYY(0),
SSYY(0),SXX2(0),SX2X2(0),SYX2(0),ACoeff(0),BCoeff(0),CCoeff(0),
X(0),Y(0), MaxLength(100),s(0),TX(0),TY(0),RV(0),ERR(0);

array: YValue[100](0),XValue[100](0);

if BarNumber = 1 then
begin
N = length;
{ be sure array size is protected }
if N > MaxLength then
N = MaxLength;
if N < 2 then
N = 2;
end;

{ fill arrays }
for value1 = 1 to N
begin
XValue[value1] = value1-1;
YValue[value1] = Price[value1-1];
end;

{ calc averages of x,y pairs }
AvgX = 0; AvgY = 0;
for value1 = 1 to N
begin
AvgX = AvgX + XValue[value1];
AvgY = AvgY + YValue[value1];
end;
if N <> 0 then
begin
AvgX = AvgX / N;
AvgY = AvgY / N;
end;

{ do regression and calc 3 coefficients }
SXX = 0; SXY = 0;SYY = 0;SXX2 = 0;SX2X2 = 0;SYX2 = 0;
for value1 = 1 to N
begin
XM = XValue[value1] - AvgX;
YM = YValue[value1] - AvgY;
XM2 = XValue[value1]*XValue[value1] - AvgX*AvgX;
SXX = SXX + XM*XM;
SXY = SXY + XM*YM;
SYY = SYY + YM*YM;
SXX2 = SXX2 + XM*XM2;
SX2X2 = SX2X2 + XM2*XM2;
SYX2 = SYX2 + YM*XM2;
end;

value99 = SXX*SX2X2 - SXX2*SXX2;
if value99 <> 0 then
begin
BCoeff = ( SXY*SX2X2 - SYX2*SXX2 ) / value99 ;
CCoeff = ( SXX*SYX2 - SXX2*SXY ) / value99;
end;

ACoeff = AvgY - BCoeff*AvgX - CCoeff*AvgX*AvgX;

{calc estimated price for desired bar
DesiredBar: use + for past and - for future
eg., for next bar in future:
DesiredBar = -1 }
oNLRPrice = ACoeff + BCoeff*DesiredBar+
CCoeff*DesiredBar*DesiredBar;

{calc std dev}
S = 0; AvgY = 0;
for value1 = 1 to N
begin
TY = YValue[value1];
TX = XValue[value1];
RV = ACoeff + BCoeff*XValue[value1]+
CCoeff*XValue[value1]*XValue[value1];
ERR = TY - RV;
AvgY = AvgY + TY;
S = S + ERR*ERR;
end;
if N-1 <> 0 then
if S /(N-1) > 0 then
S = SquareRoot(S /(N-1));

oStdDev = s;

NLR = 1;





Type : Indicator, Name : NLR1 Indicator

{ NLR1 Indicator WAV 11/26/03

NonLinear Regression Analysis
y = A + B*X + C*X*X (2nd order)

y price
x bar (time)
A,B,C coefficients to determine

modified 11/29
utilize NLR function
added price input
removed mult input
plots only last bar from curve

}
InPuts:
Length(20), price(close);

vars: S(0),oNLRPrice(0);

value1 = NLR(length, price, 0, oNLRPrice, S);

plot1(oNLRPrice, "NLR");

 





Source: https://www.tradestation.com

 

View similar (indicator for TradeStation):

Pivot Bar
Vertical Horizontal Filter
LBR_3/10 Oscillator
Optimal Tracking Filter
Draw Linear Regression Lines For Two Past Periods
Donchian Channel
Weekly Hi-Lo Lines
Moving Beyond the Closing Price
Bond Futures LR Surfing Indicator
VIDYA Variation
...
 
 
all formulas for TradeStation
all formulas

 

 

Email to friend

Top

 

     
However we try to maintain hiqhest possible level of service - most formulas, oscillators, indicators
and systems are submitted by anonymous users.
Therefore S4T™ does not take any responsibility for it's quality.
If you use any of this information, use it at your own risk. You are responsible for your own trading decisions.
Be sure to verify that any information you see on these pages is correct, and is applicable to your particular trade.
In no case will S4T™ be responsible for your trading gains or losses.
 
Privacy note | (c) copyrights systems4trading.com 2006-2012