
What is the fomula for RSI
Last post 02272007, 6:35 by wabbit. 18 replies.

01202007, 17:33 

billsingh



Joined on 01202007


Posts 23



What is the fomula for RSI
I want to scan the stock based on 2 days RSI (not 14 days) if the RSI is below 10 also if RSI above 90. What will be the formula ??
Thanks in advance
Bill



01212007, 4:44 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
The formula should be: (RSI(C,2)<10) OR (RSI(C,2)>90).
But in explorer, RSI for 2 or 3 days alway get error. It seem that in the smallest no. of days to be used must be 4.



01212007, 10:29 

billsingh



Joined on 01202007


Posts 23



Re: What is the fomula for RSI
Thanks Wblam, I will keep in mind. I want to add this scan with RMO. Lets see what happens.



01212007, 11:14 

billsingh



Joined on 01202007


Posts 23



Re: What is the fomula for RSI
I read this article http://www.tradingmarkets.com/.site/stocks/commentary/editorial/2PeriodRSI.cfm They say 2day RSI Produte beter result then 14 day.
So I ran the following.
a:=(Fml("Rahul Mohindar Osc (RMO)") > 0) AND
(Fml("SwingTrd 2")>0) AND
(Fml("SwingTrd 2")>Fml("SwingTrd 3")) AND (RSI(C,2)<10);
a=1 AND Ref(a,1)=0
and Found following matches that return 1.
AAPL CRAY GIGA RADS VOCS
I am going to watch those stocks in next few days and see what happens.
Is my formula right ???
Thanks
Bill



01212007, 11:26 

billsingh



Joined on 01202007


Posts 23



Re: What is the fomula for RSI
WBLIM you are right. 2 day RSI does not work, please ignore my last post. Here are the results of 4day RSI (just RSI)
Security Name RSI Ticker Symbol Location FASTENAL ORD 1.0000 FAST Forgent Networks, Inc. 1.0000 FORG GARMIN ORD 1.0000 GRMN SANDISK ORD 1.0000 SNDK WHOLE FOODS ORD 1.0000 WFMI YAHOO ORD 1.0000 YHOO
Lets see what happens to those stocks in next few day.



01212007, 14:18 

*PP



Joined on 04012006

Romania

Posts 141



Re: What is the fomula for RSI
billsingh:
WBLIM you are right. 2 day RSI does not work, please ignore....
Lets see what happens to those stocks in next few day.
Everything is so magic, right? ;)



01212007, 23:22 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
Hi Bill, I found a solution, create a custom indicator for 2 days RSI, call it RSI2. The formula is as follow:
U:=If(C>Ref(C,1),CRef(C,1),0)+If(Ref(C,1)>Ref(C,2),Ref(C,1)Ref(C,2),0); D:=If(C<Ref(C,1),Ref(C,1)C,0)+If(Ref(C,1)<Ref(C,2),Ref(C,2)Ref(C,1),0); u/(u+d)*100;
Then in explorer, Filer tab:
Fml("RSI2")<10 OR Fml("RSI2")>90
BTW, does RMO trading system help you make money so far? I do not have ver 10 yet.



01222007, 9:46 

uasish



Joined on 08132005


Posts 173



Re: What is the fomula for RSI
Wblam,
As in the book " Technical Analysis A to Z " by Steven B Achelis the calculation for RSI = 100  ( 100 / (1 + U/D)),where U = Avg upward price change & D = Avg downward price change,hence your code needs a modification at last line.
Asish
Asish



01222007, 10:37 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
uasish:
Wblam,
As in the book " Technical Analysis A to Z " by Steven B Achelis the calculation for RSI = 100  ( 100 / (1 + U/D)),where U = Avg upward price change & D = Avg downward price change,hence your code needs a modification at last line.
Asish
Hi Asish,
My formula is correct. The definition of the variables U and D in my formula are different from the book. In the book, U = Avg upward price change , but in my formula, U = Total upward price change. Similarly for D.
OK let me change my variables from U to A and D to V.
Now, U=A/2 and D=V/2, and
RSI = 100  ( 100 / (1 + U/D))
=A/(A+V)*100.



01222007, 17:18 

billsingh



Joined on 01202007


Posts 23



Re: What is the fomula for RSI
Wblam, Thanks for that formula, Yes I have metastock 10. Actually just strat looking into technicals. Please let me know if you want to test anything on metastock 10.
Regards
Bill



01222007, 20:32 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
billsingh:
Wblam, Thanks for that formula, Yes I have metastock 10. Actually just strat looking into technicals. Please let me know if you want to test anything on metastock 10.
Regards
Bill
Hi Bill,
I have ver 9 and thinking whether want to upgrade to ver 10. As I know, ver 10 not much new thing compare to ver 9 except a new trading system RMO system which claimed is a very successful system and help many people made a lot of money. I like to know your opinon on that system.



01232007, 2:15 

Spaceant



Joined on 08162005


Posts 189



Re: What is the fomula for RSI
wblam:
U:=If(C>Ref(C,1),CRef(C,1),0)+If(Ref(C,1)>Ref(C,2),Ref(C,1)Ref(C,2),0); D:=If(C<Ref(C,1),Ref(C,1)C,0)+If(Ref(C,1)<Ref(C,2),Ref(C,2)Ref(C,1),0); u/(u+d)*100;
It doesn't seem right with the above formula. When the above code plotted on a local index, it shows 100 and 0 for some periods.
I also plot RSI(2) on another inner box, the above code doesn't inline with RSI(2)...... seems not right to me or ....?
Sa



01232007, 6:16 

uasish



Joined on 08132005


Posts 173



Re: What is the fomula for RSI
Wblam,
If you have a particular objective of getting total upward price change then it is ok,but the link you provided where the RSI  2 periods success rate for very short time frame is very encouraging,as per their findings & the basic concept of this leading indicator lies in it's avg of up/dn .However this 2 period time frame will mostly mis match with the chart & exploration values in MS .However i am using Jose's following code ( with little tweaking for my purpose) since long.It may be of some use,the following is his original :
.........................................8.................................................................
{ Smoothed RSI indicator v1.1 } { Internal sineweighted smoothing } { Automatic historicallyvalid RSI peak/trough boundaries} { Boundary crossover signals: +1=Long, 1=Short} { ©Copyright 2004 Jose Silva } { http://www.metastocktools.com }
{ user input } pds:=1/Input("RSI periods",1,2520,10); plot:=Input("plot: [1]RSI, [2]Long/Short signals",1,2,1);
{ RSI up/down average components } x:=If(C>Ref(C,1),CRef(C,1),0); y:=If(C<Ref(C,1),Ref(C,1)C,0); up:=PREV*(1pds)+x*pds; dw:=PREV*(1pds)+y*pds;
{ sineweighted internal smoothing } s1:=Sin(30)*up; s2:=Sin(60)*Ref(up,1); s3:=Sin(90)*Ref(up,2); s4:=Sin(60)*Ref(up,3); s5:=Sin(30)*Ref(up,4); up:=(s1+s2+s3+s4+s5)/(Sin(30)*2+Sin(60)*2 +Sin(90)); S1:=Sin(30)*dw; S2:=Sin(60)*Ref(dw,1); S3:=Sin(90)*Ref(dw,2); S4:=Sin(60)*Ref(dw,3); S5:=Sin(30)*Ref(dw,4); dw:=(s1+s2+s3+s4+s5)/(Sin(30)*2+Sin(60)*2 +Sin(90));
{ RSI indicator } dw:=If(dw=0,.000001,dw); RS:=100100/(1+up/dw);
{ automatic RSI peak/trough boundaries } pk:=Ref(RS,1)>Ref(RS,2) AND Ref(RS,1)>RS AND Alert(Ref(C,1)>Ref(C,2) AND Ref(C,1)>C,2) {AND Ref(RS,1)>50}; pkVal:=ValueWhen(1,pk,Ref(RS,1)); pkAvg:=Cum(pkVal)/Cum(pkVal>1);
tr:=Ref(RS,1)<Ref(RS,2) AND Ref(RS,1)<RS AND Alert(Ref(C,1)<Ref(C,2) AND Ref(C,1)<C,2) {AND Ref(RS,1)<50}; trVal:=ValueWhen(1,tr,Ref(RS,1)); trAvg:=Cum(trVal)/Cum(trVal>1);
{ RSI peak/trough boundary cross signals } up:=Cross(trAvg,RS); dw:=Cross(RS,pkAvg); signals:=updw;
{ alternative RSI peak/trough signals } {up:=Ref(RS,1)<Ref(RS,2) AND Ref(RS,1)<RS AND Ref(RS,1)<trAvg; dw:=Ref(RS,1)>Ref(RS,2) AND Ref(RS,1)>RS AND Ref(RS,1)>pkAvg; signals:=updw;}
{ plot } If(plot=1,pkAvg,0); If(plot=1,trAvg,0); If(plot=1,RS,signals)
..........................................................8......................................
Asish
Asish



01232007, 11:08 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
Spaceant: wblam:
U:=If(C>Ref(C,1),CRef(C,1),0)+If(Ref(C,1)>Ref(C,2),Ref(C,1)Ref(C,2),0); D:=If(C<Ref(C,1),Ref(C,1)C,0)+If(Ref(C,1)<Ref(C,2),Ref(C,2)Ref(C,1),0); u/(u+d)*100;
It doesn't seem right with the above formula. When the above code plotted on a local index, it shows 100 and 0 for some periods.
I also plot RSI(2) on another inner box, the above code doesn't inline with RSI(2)...... seems not right to me or ....?
Sa
I found out the reason. MetaStock uses Wilder's smoothing average. My formula is equivalent to using simple average.
You may try this:
A:=If(C>Ref(C,1),CRef(C,1),0); B:=If(C<Ref(C,1),Ref(C,1)C,0); u:=Wilders(A,2); d:=Wilders(B,2); u/(u+d)*100;
It should be looked the same now.



01232007, 11:22 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
uasish:
Wblam,
If you have a particular objective of getting total upward price change then it is ok .....
Mathematically, if simple average is used,
U=A/n and D=B/n
then U/D = A/B.
uasish:
... the link you provided ...
The link is not provided by me, is by Bill.



01232007, 11:54 

billsingh



Joined on 01202007


Posts 23



Re: What is the fomula for RSI
Wblam, I recommend upgrade and yes I know few people who have 10 consective winning trade using RMO what is what made me buy Metastock ver 10. For this year in Jan it made me over 20k with one trade and I am in the second trade with 5k up using RMO.
Go get it tiger, it is worth the money you pay for upgrade.
Regards
Bill



02242007, 17:51 

wabbit



Joined on 10282004

Perth, Western Australia

Posts 3,074



Re: What is the fomula for RSI
Bill and All, You don't NEED to upgrade to MS10 to get the RMO as it is exactly the same as the Rainbow Oscillator that was included with previous versions of MS (as early as MS7). There are other good reasons to upgrade to MS10, besides the RMO. wabbit
Please update your thread status:FREE Equis Formula Primer : http://forum.equis.com/files/19673/download.aspxTo post code, *code* your code */code*To post links, URL goes here For custom MetaStock and DLL programming : http://www.wabbit.com.auContact via SkyPE



02242007, 23:50 

wblam



Joined on 11252006


Posts 79



Re: What is the fomula for RSI
Hi Wabbit,
Thank you for the information. I found the Rainbow Oscilator indicator. However it is not documented in the manual of ver. 9.1.
Can you teach us how to create a template like the one in ver. 10?
Thanks.
Lam



02272007, 6:35 

