Correct me if I am wrong suppose there are 2 correlated stock pairs i.e if one stocks goes up x % other goes up by approx x %. Now here is a strategy if one of the stock fails to go to x % it will try to go up x% meaning i will buy a call of this stock and the other stock which has gone x % I will buy a put. Meaning the stock that underperformed the other stock will try to go up and the stock that overperformed will come down to the other stock price. Hence, I should make profit. In worst case if the 1 stock moves up the other will move up too meaning the profit made by the call covers the loss made by the put of the stock 1. Is this a solid strategy assuming the 2 stocks are very correlated.
its called pair trading
It looks like the validity of this strategy depends on how valid this assumption about the 2 stocks is.
How, valid is this assumption (going to be) ? 
1 Like
Exactly… this often happen with stocks with high weightage in any broad indices. BEWARE of any blackswan though.
Through what i have seen using RVNL and IRFC show a very good correlation also RECLTD and PFC show correlation. Lets say they are correlated about 70% of time then do u think this strategy can be profitable ?.
Even with that assumption,
the strategy being profitable in absolute terms (i.e. even beating hard-cash at rest),
depends on certain aspects like…
a. Proper bet-sizing to ensure one’s capital isn’t wiped-out by an unfavorable streak early-on.
Checkout Gambler's fallacy - Wikipedia.
b. The thresholds of profit-booking and stop-loss would one set to come out net-positive even with a 70%-30% biased outcome. What would the thresholds need to be, to achieve a net-positive expected-value upon one’s initial investment? If you share the math, other interested folks could review/share their thoughts on it. Remember to account for overheads involved in active trading.
Also, even with all this effort involved,
being profitable relative to other lower-risk/zero-risk investments,
is another aspect that we would need to consider
before deciding to deploy one’s presumably limited time and capital into this strategy,
(…or opt to deploy the limited capital in alternative investments with less uncertainty and other risks, and/or higher returns, and ideally ones that leave more time free for one to pursue other value-generating ventures.)
So you are saying that i should first backtest with different params like size of position, the tp and sl etc and check if it beats 12% if not just buy mutual funds.
Well, yes.
We test as much as we can, as soon as we can,
and ideally, risking as little of our capital as we can.
If we find that historically some combination has worked in the past,
can then evaluate whether the current/upcoming conditions are similar enough now
that similar behavior might repeat again.
And even then if they are likely to repeat,
if the likely returns from the scheme aren’t going to exceed
some other far more diverse / less risky / simpler alternatives available,
then sounds prudent to just stick to the alternatives instead. 
Can a proposed scheme beat 4-5% return (after taxes) ?
If not, then T-Bills, Liquid-ETFs, DICGC-insured bank FDs, …
might be better alternatives depending on the liqudity-profile one desires.
Nope.
Correlation means nothing and past is not a guidance for future. Also the duration of divergence / convergence is non-deterministic. Everything is correlated that way to some extent or the other. What you need is Co-Integration. Co-Integration should be first proven statistically using various tests over long period of time. And as it turns out - it is very rare among two different securities even in the same sector/industry/group of industries. Typically exists momentarily for the same underlying and then vanishes due to arbitrage. Very hard to get more than fixed income returns. Check any arbitrage fund.
Buy hey … You should try … And check it yourself.
1 Like
That’s what I found out — that I should use cointegration. I backtested it on around 3 years of data, and it gave only about a 60% return, which is kinda bad compared to my own swing trading strategy. That one takes way less effort and fewer trades than the cointegration strategy. But still, I’m trying to improve it on weekends. Let’s see how far I get. Even if it turns out to be a bad strategy, at least I’m learning some of the math behind the stock market and picking up Python too — so it’s still a win.
1 Like