Zero sample correlation does not necessarily mean there is no linear relationship - page 60
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
For some reason I had a coffee for the night and had to develop a system for prototyping statistical arb. strategies based on my model, because the pile of testing scripts has grown to unimaginable size. Now any pair can be tested and put into real trading in a couple of minutes (the trading part has been written and running for a long time).
I will show off, perhaps, some nice pictures. The labels that can be used to fully or partially reconstruct my approach are painted, of course. The testing is done with all costs in mind.
Explain if you can.
The very first picture. I understand that both assets have been growing for the whole 1096 days, rather monotonically?
Explain if you can.
The very first picture. I understand that both assets have been growing for the whole 1096 days, rather monotonically?
Right. What's so surprising about that? They are shares in the same company, they are cointegrated and therefore grow together.
There's something wrong with the traders count, the sum doesn't match the longs and shorts.
Everything is correct there :) Longs and shorts are trades that led to the opening of a position in a specific direction (this is more convenient for me); closing can take place without a position reversal - then only the total is increased, while longs and shorts remain the same.
This is quite obvious. It is not at all obvious the origin of such a significant number of shorts (67/98=68%) on a monotonically growing asset.
Longs and shorts count towards the portfolio. The specifics of the strategy are that we always sell one stock and use the money to buy another. Longs and shorts only differ in terms of which stock we sell and which we buy.
Also, monotonic growth in a stock does not tell us anything, as the value of the portfolio changes quite differently.
Longs and shorts count towards the portfolio. The specifics of the strategy are that we always sell one stock and use the proceeds to buy another. Longs and shorts on the portfolio differ only in which stock we sell and which we buy.
Also, a monotonic rise in a stock says nothing, as the value of the portfolio changes in a completely different way.
I think I got it. Next.
The real asset fluctuates around the virtual synthetic: if the asset is above it, then short, if the asset is below, then long. As soon as the deviation from the real synthetic has become zero, then exit. Right?