I'm sure every CL trader longs for those days when it regularly moved two handles and better during RTH. Yet, even when the daily range is more or less the same as ES, some traders have learned and become comfortable with CL and the move to ES (or for many/most of us the move back to ES) holds little or no appeal. And, of course, every CL trader believes those two and three handle days will be back; whether that is rational or not!
If You Trade One Instrument Only, How Do You Deal With This?
- Thread starter tradingjournals
- Start date
Quote from SimpleTrades:
If you're calculating variance, you're doing so on preexisting data. The trade(s) are already done.
I understand from your original post , that it does not make a difference whether one trade one or multiple strategies. So I asked you what about the variance?
Trading CL is perfectly fine... just like any other futures symbol of personal choice. But the days of CL being far superior to all else are over with, forever. Unless the Dodd-Frank Act laws of large-spec limits are redacted.
The term "comfortable" is relative. Comfortable to do what? Trade one CL contract? Five? Twenty? I don't know anybody who dares trade block size CL any more, most people dick around with one contract all day long taking a bazillion turns hoping that adds up to something.
Eventually a trader matures above the stage of churning and realizes it is easier, much more lucrative and a lot more lifestyle pleasurable to size-up contracts, take a few trades in the morning, book overall larger $$ gains and quit work for the day. That is why all of us began trading in the first place: lifestyle improvement.
The only true way to keep increasing rates of success is adding contract size to trades. There is a concrete ceiling above as to how many ticks per day anyone can average, and that is a low ceiling for everyone with zero exceptions. So unless one is comfortable adding lots of size over time to CL trades, it is a self-limiting act. Nothing wrong with that, if you don't mind limiting yourself to lesser annual profit potential.
What I ended up doing to compensate for this in my backtesting is subtract a tick of profit and add a tick of loss to every entry or exit that involved a stop order. Sure there are many times when stops are filled at the stop price, but the extra slippage costs I padded into my analysis of static charts covered those rare live trades when the slippage was 3 ticks or more on CL.
Consider a trivially simple example. Buy 500 shares of a theoretical instrument for $1, buy 100 shares of that same instrument for $2. Now, reduce it to its simplest form. How many shares have you actually bought, and for what price? Despite the fact that you have carried out two trades, you can actually express those same two trades as only one theoretical trade. You can apply this same principle to many different instruments, prices and volumes.
Realizing this, can you see how variance is actually built right into the weighted sum? At the end of the day, you have traded one theoretical instrument, and you're either positive, negative or flat.
If V=Var(A)=Var(B), Var (a/2+b/2)= V/2 +1/2Cov(A,B).
Cov(A,B) could be a great friend.
If there is a large number of vehicles with capital split equally among them and with equal variance, the variance tends towards the average correlation coefficient times the original variance (and not the original variance as you seem to think).
Also if you have ever wondered why a moving average looks smooth, now you should know why.