Cumulative RSI 2 periods strategy

Cumulative RSI 2 periods strategy

Long only strategy that performs extremely well on any world indices and beyond on a daily timeframe. I will certainly do much exploration of this very short term mean reversion strategy, as it seems very powerful without doing any change on triggers. 

The main trigger is based of a “cumulated RSI” over the 2 recent periods while trade are launch only if the price Close is above a 200 period moving average.

When the cumulated RSI enter in oversold territory, we expect the price to return to its mean, on the bullish trend. We exit the market when the CRSI gain overbought area above the 65 level.

As for my current test : (with exactly the same code on CFDs)

CAC40 :  (featuring picture) : 10€ /contracts, 1 contract per trade since 1988 = +370% / % Drawdown max = 12.7%

IBEX35 : 2€ / contract, 1 contract per trade since 1994 = +134.33% / % Drawdown max =  25.6%

SP500 : 1€ / contract, 10 contract per trade since 1984 = +95.33% / % Drawdown max = 8.08%



Share this

Risk disclosure:

No information on this site is investment advice or a solicitation to buy or sell any financial instrument. Past performance is not indicative of future results. Trading may expose you to risk of loss greater than your deposits and is only suitable for experienced investors who have sufficient financial means to bear such risk.

ProRealTime ITF files and other attachments : How to import ITF files into ProRealTime platform?

New! PRC is also now on YouTube, subscribe to our channel for exclusive content and tutorials

  1. TradingDeckard • 02/17/2016 #

    Very interesting strategy and easy one! Have you tested it in any other markets than your examples??

    • Nicolas • 02/17/2016 #

      Not yet. A nice thing to do would be to implement another strategy into this one when price go bearish and dive under the MA200. Of course, it works well because of the mean reverting thing over a nice bullish movement, but adding other strategy over this one would be great, diversification is the key. As i said in this post, i need to work further on it, you are right, it is very interesting!

  2. f.favret • 02/17/2016 #

    Avec une période de 2, la formule “CUMRSI = SUMMATION[Period](RSI[Period](close))\"
    équivaut elle bien à
    \"CUMRSI = RSI[2](close de la barre courante) + RSI[2](close de la barre précédente)\"?

    • Nicolas • 02/17/2016 #

      FR/ Bonjour f.favret, en effet c’est bien la même chose!
      EN/ CUMRSI on 2 period is the same thing as additional of last 2 RSI values :
      // example of the same result

      r = RSI[2](close)

      result = r+r[1]
      CUMRSI = SUMMATION[2](r)

      RETURN result as \"result\",CUMRSI as \"cumrsi\"

    • f.favret • 02/17/2016 #

      Merci / Thanks!

    • f.favret • 02/17/2016 #

      Why not summation of the 2 lasts RSI 14 period ?of  other period?

    • Nicolas • 02/17/2016 #

      A lot of quantitative traders believe that the RSI default ’14’ periods is dead, although for what it were developed to : spot the overbought and oversold areas of the price development over time. It does not mean that the RSI formula is mean-less, but not still effective with its default value for whom it were set, back in 1978 by M.Wilder! The cumulative RSI is nothing more than another indicator, it borrows the formula from RSI, cumulate 2 periods and that’s all. Why 2 periods for this one? because price behaviour, return to the mean fastly, no more no less.
      Automatic trading and quant strategy development are a vast playground, go on, play with period, create another indicator from this one or a fresh new one, if it works you were right, if not you still have an entire life to learn maths 🙂

  3. Doctrading • 02/17/2016 #

    Very nice and logical strategy…
    Could be used for example to take orders on the french “PEA” (Plan d’Epargne Actions).
    “RSI 14” is still alive !
    Best regards 😉

  4. Stef • 02/17/2016 #

    “1 contract per trade since 1988 = +370%” – is this per annum, or over the period? How much was your starting capital?

  5. Nicolas • 02/17/2016 #

    Over the whole period. I think that the starting capital tested were 10k€, this is my default value in the ProBacktest.

  6. gianlox • 02/17/2016 #

    I created the screener of this TradingSystem  but it looks like repainting, if today the screener gives me the signal to buy, the next day the TS does not draw anything on the chart. 

    • Nicolas • 02/17/2016 #

      Maybe because you screen with EOD data? So in this case you are already late.

  7. ML • 02/17/2016 #

    Hi All! Have you succeeded in developing a strategy also for the bearish side of the market? I tried but without success…
    Thanks a lot!

  8. jctrader • 02/17/2016 #

    Bonjour .
    CAC 40    1988 = 1000   today = 4900  soit 490%  en buy and hold  ….
    + 370% avec cette stratégie : désolé je ne saisis pas bien l’intérêt .

    • Nicolas • 02/17/2016 #

      Aucune immobilisation du capital. Quel est le drawdown du buy & hold ? Je ne l’ai pas calculé moi même, mais cela a dut être difficile en 2008 et 2012 ? Après coup, lorsque l’on a toutes les informations à disposition, certaines stratégies auraient été meilleures que d’autres. Il est vrai que l’on a coutume de comparer les stratégies d’interventions automatique avec le buy&hold sur les indices et les actions, malgré cela, une stratégie qui fait 370% à elle seule et qui peut être incorporé à un portefeuille plus vaste de stratégies automatiques est valable selon moi.
      Cette stratégie de mean reversion est tellement simple, qu’il aurait été dommage de ne pas la proposer à tous 🙂

    • jctrader • 02/17/2016 #

      ok pour le codage mais le choix “indice” n’est pas le bon : moins de 5%/an pour le meilleur et pour l’IBEX 1% annuel avec 25% de DD …..

Register or


avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar
Related users ' posts
gabri GraHaL, I cannot see any link attached. I mostly manual trade but for stocks with strong fi...
GraHal Yes sorry, I set up a link to a screen shot on my google drive and then I got locked into th...
gabri Here's the thread
nwesterhuijs Thanks, only saw it just now.
juanj For the latest version and discussions see the Ichimoku thread here: https://www.prorealcod...
Louwrens Hi Juanj. Thanks for this. I am tying it as we speak. It does not trade that often, which is...
diegofe_2000 cordial saludo NICOLAS , gracias por excelente indicador. ¿ es posible obtenerlo en MT4 ? ...
Nicolas Lo siento, pero no estoy trabajando para ayudar a la gente aquí de forma gratuita para promo...
diegofe_2000 Nicolas , gracias por tu atención. La pregunta era por curiosidad (yo no trabajo ni manejo ...
9 months ago
CSR strategy DAX 1 D
CSR strategy DAX 1 D
ALE Thanks as always Nicolas
Toto le Heros It looks very interresting. I have one question related to the following result in ProbackTe...
StefanStrand I tried running this on IG ProRealTime and get the following error message, why could this b...
9 months ago
Francesco78 very nice, thanks!
finplus Thanks for the job. Which variables do you suggest for timeframe 1 hour? 
Maz Depends massively on your market and the volatility. I suggest using the variable optimizer ...
10 months ago
HerveS_67 Bonjour, A quoi correspondent les données S, M, L ? Hello, What do the data S, M, L?
gabri Excellent job like always!! are you working also on the Barycenter indicator?
Nicolas No but I think you can find a lot of different version of the "center of gravity - COG" indi...
Juan Salas Hi Francesco, Yes, this is pretty much the question. The system ask me a number of contract...
Francesco78 ok sounds good, you can put min size = 1 and go for the seasonable breakout too, also you ca...
JR1976 HI all , anyone test 200k bar for this great code ?
Nicolas All conditions under parenthesis for the c1 to c4 conditions should be inverted. 
tomus Can you give an example of the overbought codes please? Thanks.
ams123 Frank Merci Nicholas -:)
10 months ago
gabri Thanks John. I guess you can change line 2 to use some other indicators. I use only RSI and ...
JohnScher Only the indicators change in line 2? That's all? I'll try it! I'll answer when it's done. ...
Mert Thanks, Gabri.
Nicolas Another fine piece of code gabri! Excellent work again. Thanks a lot for your precious contr...
gabri Thanks Nicolas!!
bolsatrilera bravo, molto intelligente.
owes29 Do a Youtube search for "Revolutionary Indicators Reveal Tops & Bottoms" it explains a v...
owes29 Hi is there anyway to develop this for the pro scanner on daily or hourly scans. so it woul...
Nicolas Of course, please add a query in the proscreener forum.
gregus bonjour a tous quelqu un pourait il recodé en prt il sagit  du dynamique zone ma, je n arri...
Nicolas Merci de faire une requête spécifique sur le forum. 
1 year ago
rfsteve Nicolas what do you think of these modifications for 233 tick charts ?         cheers Steve....
Nicolas Thanks for this modification, I still do not have tested it, where did you get this idea to ...
rfsteve Trial and error from study of indicators call it coding mad science was trying to find an in...
David Nicolas I've looked at it again and there's a problem.  As an example here's a simple cross...
Nicolas You'll need to preload bars to get the good calculations of you indicators. I did not test i...
David Nicolas I tried DEFPARAM Preloadbars = 5000 And still the same drawn output of entries/exit...
1 year ago
Nicolas I don't know about the formula of the one built-in into the platform. This one has been conv...
verdi55 I found the reason. The indicator built into the PRT platform displays only the StochRSI var...
Nicolas Well, in this indi you can choose the RSI period, the period of the stochastic calculation o...