Yamato
Legendary member
- Messages
- 9,840
- Likes
- 246
correcting comparison study on CL systems
I have double-checked my studies on the CL systems, here:
http://www.trade2win.com/boards/trading-journals/85510-my-journal-2-a-210.html#post1450650
And there was a mistake, because in fact the two CL systems, CL_ON_2 and CL_ID_3 were not compared to the CL but to the ES, which stayed there from the previous study on ES_ON_2.
I will quote the previous studies here and re-analyze the whole thing. On top of it, I will add two more systems which I would like to suggest for inclusion in the traded basket: CL_ON_3 and CL_ID_4.
(Incidentally, that one xanax pill I took two days ago is still giving me some amnesia, so that I have been forgetting to write words in emails and posts).
PREVIOUS (WRONG) STUDY SAID:
Without even reading it, I will first write the new study (including the two new CL systems:
CL_ON_2 correl 0.68
CL_ID_3 correl 0.76
CL_ON_3 correl 0.77
CL_ID_4 correl 0.45
Ok, the first three have a correlation with where the CL goes, but each one of them falls less when the CL falls, and that is why they're worth trading. In a descending order of performance, their capability of not falling when the CL falls goes like this:
CL_ID_3
CL_ON_3
CL_ON_2
On the other hand, the CL_ID_4 actually makes more money when the CL falls, so they should all be traded in a balance like this: the three "bad" ones should be allocated 1 contract, while the good should be allocated 3 contracts. And this is just my CL portfolio, because if I do something like this, I'd need to allocate 3 more contracts to all other systems in my portfolio.
Right now we're only trading half a contract on CL_ID_3 and CL_ON_2 so I won't bother the investors to ask them and make any changes, but next time we'll talk about selecting systems to add, I will make sure to tell them that we should include CL_ID_4 at least with a half contract. Also, CL_ON_3 should be added, but then if we do this, we should have 1 full contract on CL_ID_4.
At the moment of choosing our new wave of systems, last week, I did not ask to include the CL_ID_4 because on Monday it would have been hit (but there was no trading because it was a half holiday) very badly by that huge >5% move in CL (Libyan problems and stuff). I got very scared, and chose to forget about it, but in fact it is the best one of all in terms of back-tested health, in that it does not suffer from CL's falls.
I have double-checked my studies on the CL systems, here:
http://www.trade2win.com/boards/trading-journals/85510-my-journal-2-a-210.html#post1450650
And there was a mistake, because in fact the two CL systems, CL_ON_2 and CL_ID_3 were not compared to the CL but to the ES, which stayed there from the previous study on ES_ON_2.
I will quote the previous studies here and re-analyze the whole thing. On top of it, I will add two more systems which I would like to suggest for inclusion in the traded basket: CL_ON_3 and CL_ID_4.
(Incidentally, that one xanax pill I took two days ago is still giving me some amnesia, so that I have been forgetting to write words in emails and posts).
PREVIOUS (WRONG) STUDY SAID:
7) CL_ON_2 correl 0.89
The most correlated of all 7 tested so far, because its correlation does not come from the fact that both rise endlessly but to the fact that it falls at the same time the CL falls. Nonetheless, it is not so disgusting after all. Sometimes the rising and falling don't match one another, which is good and also, when the CL falls, the CL_ON_2 does not fall as much, so it is still far better than just buying the future and staying long on it, without ever closing it. In fact, if you are not fooled by the two y-axis scales (one on the right and one of the left), you'll notice that CL_ON_2 makes 500% while CL makes nothing at all (from 100 to 100). 500% is not peanuts, and that is why we're trading it live. Having said this, CL_ON_2 has the strongest correlation with its underlying of all 7 systems tested so far.
8) CL_ID_3 correl 0.21
Oh yes! This is an excellent one, for both overall profit (1200%) and lack of correlation. Just see it to believe it. In fact it is the best one we have and we should be allowing it 1 contract instead of just half (the small contract).
Without even reading it, I will first write the new study (including the two new CL systems:
CL_ON_2 correl 0.68
CL_ID_3 correl 0.76
CL_ON_3 correl 0.77
CL_ID_4 correl 0.45
Ok, the first three have a correlation with where the CL goes, but each one of them falls less when the CL falls, and that is why they're worth trading. In a descending order of performance, their capability of not falling when the CL falls goes like this:
CL_ID_3
CL_ON_3
CL_ON_2
On the other hand, the CL_ID_4 actually makes more money when the CL falls, so they should all be traded in a balance like this: the three "bad" ones should be allocated 1 contract, while the good should be allocated 3 contracts. And this is just my CL portfolio, because if I do something like this, I'd need to allocate 3 more contracts to all other systems in my portfolio.
Right now we're only trading half a contract on CL_ID_3 and CL_ON_2 so I won't bother the investors to ask them and make any changes, but next time we'll talk about selecting systems to add, I will make sure to tell them that we should include CL_ID_4 at least with a half contract. Also, CL_ON_3 should be added, but then if we do this, we should have 1 full contract on CL_ID_4.
At the moment of choosing our new wave of systems, last week, I did not ask to include the CL_ID_4 because on Monday it would have been hit (but there was no trading because it was a half holiday) very badly by that huge >5% move in CL (Libyan problems and stuff). I got very scared, and chose to forget about it, but in fact it is the best one of all in terms of back-tested health, in that it does not suffer from CL's falls.
Last edited: