I made a good honest effort, investigating it in depth. I too was looking for shorter term mean reversion strategies off the shorter term cointegration (days/weeks) approximately 2 years ago. The conclusion I came to is that the definition that cointegration is a long term co-movement is accurate. But I’ve read some had success with this. Drop a line if you get any leads that look promising.
I’m still testing cointegration with forex. So far, it seems to be better with larger timeframes, as FXEZ said. On smaller timeframes the equations don’t work very good.
But also I’m still testing kelton’s method. I think we have a winner, holy grail like system if we use good money management.
Just check my stats: Kelton’s Method System | Myfxbook
trading since August I had no losing month, with an average profit of 12.27 % monthly. I think that results far exceeds the market, and if compounded can transform a little amount (1000 USD) to a very vast wealth (16.5 million) in 7 years, maybe less with a higher risk trading.
I’m still testing the statarb with cointegration. I found that on 15m it doesn’t work. Now I’m testing on 1h, it is somewhat slow, but it seems to work better. Too soon to conclude anything yet, but I’m forward testing 
Meanwhile I’m also testing Kelton’s system, with the modifications I did about stop loss and take profits. It is working pretty good I started my real account on July 12, since then I have a 52 % profit or about 10 % monthly. I already did 214 trades, only 7 of them were losers. If you check the myfxbook stats, the system have a 58 % winners vs 42 % losers, and the average win is 83.25 pips vs average loss of -73.85 pips.
That leaves with a system that wins most of the time, and when it wins, the profit is bigger than when it lose. With a positive profit factor of 1.59, I can almost conclude that this system is in reallity a holy grial I’m going to test it more, until end of the year, after that I think I will go a little bigger on it.
Thanks to Kelton for this illustrative example of trading
Getting good results there Medi, thanks for the update!
I’ve been too busy to do much lately.
Hey everyone, thanks so much for giving my strategy a chance. Im glad to see you all have profited and or tweaked it to fit your needs. I am starting an Instagram page and I would love for everyone to follow me! I post everything about stocks, currencies, and anything else about the financial markets!
Instagram name is – investmentbrothers
-Kelton Williams
Hey Medisoft, thanks so much for your support. This was the best reply to my thread I have read. Good luck and keep me updated on your success!!
Follow me on face book or instagram!
my full name is Kelton Williams and I live in Virginia and Im 21
my instagram is investmentbrothers
Thanks again!
Thanks to you
I’m following now on instagram
Perfect Man! Hit me up on facebook if you have one!
Hey Everyone, I just wanted to thank you all for your support and contribution to this strategy to make it everything it can be. I just wanted to post some new findings that might make everyone here some more money
Now as we all know that the charts of course if you fixed the scales they could never touch again and you would be either in the money or out of the money with no real way of knowing when you get out.
And on the other hand if you don’t fix scale the charts they could move " mid-trade" and one could currency that was on top is now on bottom and you still haven’t made profit.
Try this…
[B]#1 - Use the Vitrite method if possible
#2 - Do not fix the charts
#3 - Wait for the divergence to be around 30-50 pips
#4 - Trade until the two currency charts close.[/B]
Now what I have found is that if you wait for a large enough divergence that the charts wont shift enough that they touch and your still out of money.
Even if they touch and you make money but not as much as you thought because of the shift close the trade.
If you are out a little money and they touch because of the shift ([I] Happens rarely [/I]) close the trade and take the loss.
Its been working for me with no losing trades so far, but with crystal clear entries and exits.
Let me know what you find, and again im open to all and any criticism.
Thanks so much again!
Find and add me on FB or Instagram too!
FB - Search Kelton Williams in fairfax Va
Instagram - Search investmentbrothers
Best Regards,
Kelton W.
Wtf?
What is Vitrite? I still can’t get it…
Hey Henlatourrette, Vitrite is a computer program that allows you to make one of your computer windows translucent so you can see behind it. My Vitrite method to seeing the correlation between currencies is to open up two seperate MT4 programs and put the windows on top of each other and then make one of the windows see through by using vitrite. So you can see the two charts overlapped without an indicator. A short description should also be on the first or second page along with a picture of what the method looks like.
Hi All, and particularly Medisoft, Jedstar & all that have further developed Kelton’s method and are trading it with success!
I just stumbled across this thread now that has 132 pages and I tried to read it all, which turns out to be quite an undertaking! It seems that there has been quite some development in various methods, with a lot of adjustments particularly in the area of balancing and risk management.
I am still determined to read all but I am still wondering if those that are trading this method successfully could summarise their current way of trading Kelton’s strategy (entry, exit, stop loss, used indicators, etc…). This would be very helpful and highly appreciated!! I am planning to seriously test this method myself and I am looking forward to deliver valuable contributions to this thread as well!
Thanks a lot!
Hey Pipsi its great to meet a new follower!
I personally have not been actively trading this strategy I made just because it takes time to watch the markets and I have a very
busy full time job that limits my screen time.
On the other hand you should also check out my other strategy I posted called ( Keltons new 40% strategy ) It can be used with a stop so no need to have to watch the markets all the time… just manage them. Also easy to use on a mobile device.
Thanks for support my thread though! I hope you profit like me and so many others from it!!
Hi Guys,
It has been a while since I posted on this thread. I always really liked the simplicity of this strategy, but my issue was the potentially large drawdown, whilst in the pursuit of relatively speaking, little profit. The correct way of course to manage this was to properly manage the risk and set appropriate lot sizes. Unfortunately, risking say 1% per trade with stops in excess of 400+ pips gave very little profit.
I changed the way I was looking at it slightly, by trying to trade with a smaller SL based around how many standard deviations of movement there was against (or with) a trend. I’ll be honest and say, I got a little disinterested in that idea as again during my testing I wasn’t sticking to my own rules. If I couldn’t follow my own rules in a test invironment, there was no realistic way I could trade it in a live environment…
Anyway, I’ve been following the thread on and off, and over the last few days, I’ve read it all again from the beginning. Yep, all 1,314 posts! So, a belated thanks for all the great posts. The cointegration side of things sounds really interesting, but strangely for me, even though I’m quite happy trading off the 4H or daily charts, I suspect I won’t have the patience to work with that.
Anyway, my interest has be re-kindled over the last few days, so hopefully I’ll start to contribute again, maybe with something worthwhile (and maybe-not)…
The cointegration didn’t work on 1H. I have a lot of losers a very small number of winners. I think that because the pairs have high correlation that gives spurious cointegrations.
Hi! On EU/GU I found that a stop loss of 400 pips is pretty good. You are almost never going to reach that level, so that’s good for risk calculation.
My take profits are based on the divergence I took. If I entry at a 100 pip divergence, I look for 61.8 % of that (just a fibo number hehehe).
But note that I fix my charts and measure the divergence based on that. That is different from Kelton’s method. But that is also giving me about 10 % monthly since about 6 months.
Thank you all for your responses!
Currently I am testing both Kelton’s original method (1M, no scale fix, no stop loss) as well as Medisoft’s method (Jedster’s method will be next ;-))
Kelton’s method is quite straight forward (thanks a lot, Kelton!!!).
As for Medisoft’s method, this is how I have set it up:
I open one MT4 instance with the following charts (all with Medisoft’s “keltonprimarypair” template – thanks a lot, Medisoft!!!)
- EURUSD – D1, plus an indicator window showing ATR (default setting 14)
- EURUSD – M15, plus an indicator window showing the correlation, using the indicator “Cor.mq4” with default setting 20 (can be found when searching for “mt4 indicator cor” in google
- EURUSD – M1, no additional indicators
One the second MT4 instance I have the same setup for GBPUSD but with Medisoft’s “keltonsecondarypair” template. This MT4 instance is the one I make transparent using Vitrite.
My current “routine” (can you call it routine if you have only done it a few times?):
- Check EURUSD M15 chart if correlation is 90% or more at the moment. If it isn’t I look for the last time the correlation was >90%. The price at the time then becomes the basis for the next step
- Check daily ATR for EURUSD and fix the scale of the M1 chart accordingly. Basis is current price or the price of the last time the correlation was > 90%. The maximum = price plus current ATR, the minimum = price minus current ATR.
- Do the same for GBPUSD
From then on I keep checking if prices diverge by 20 pips, measured on the MT4 instance that shows the secondary pair (GBPUSD). When entering a trade I use the Kelton EA provided by Medisoft (thanks again, Medisoft!!!). I exit a trade when it shows profit (not quite sure about take profit yet but will stick with 61.8% as per Medisoft’s previous post).
@Medisoft: Does this sound like a reasonable setup to you?
What I am particularly wondering:
- When you talk about ATRs for scale fixing or for determining the lot size, are you talking about the daily ATR with default value (14)?
- How long do you stay in a trade? Currently I am in a trade (divergence was 20 pips) for more than 12 hours and there is no real sign of prices coming closer together. They are just running pretty much parallel, 30 pips apart. As I see it I could
A) Close the trade, accept the loss, realign the charts and look for a new trade opportunity
B) Wait and hope that the prices come closer together and close it at profit
Any advice?
Thanks a lot!
There are several ways round you problems,
Trading pairs less than 85% but more than 60% correlated is better, as when very correlated pairs go out of sync they stay out for a while,
Set a time limit on the trade before you open it,
Don’t be a long wrong ( pro tip)
Also trade many pairs with smaller position sizings that way one bad divergence won’t bit your behind bad.
Trade volume through many (oops tip)
Hi Pipsi
14 days daily ATR is good. My average trade duration is 11 days, and sometimes it diverge up to 300 pips. The biggest divergence I registered is 423 pips, but it was ocurred only once.
A fact of this strategy is that when the pairs are trending you are going to see them runing in parallel for some time. One is faster than the other, so your net divergence is growing and growing. Then, they stop that trend and reverses. And the inverse process reduces the divergence and makes you profit. It could diverge on your favour for some time also and then returns to the other side. That’s the reason I take profits on the 50 % of the divergence I entered, and with fixed charts I try to enter my first basket on a 100 pip divergence, to win 50 pip.
This makes me think that I should measure the time that the pairs diverge and then converge again to know the total cycle time. If there is a pattern there it could help to reduce drawdown and so increase the lot sizes.
I’m starting right now measuring it 