4Mes
Interesting approach. Above all, the basic principle of investing based on rules is likely to be successful in the long term.
But it also depends on the system. I had already experimented with a commodity GTAA model. But then I abandoned it for various reasons. I would have the following fundamental concerns with your model:
1. it is untested. Momentum models need parameters that exactly match the assets used, otherwise they won't perform.
2. momentum only works very moderately with commodities. Not all asset classes produce classic momentum. Oil, for example, does not work well, the movements are too erratic and driven by news. Momentum gains are more of a coincidence. You could check each commodity to see what the typical momentum intervals are and then use the average to determine the parameters. But I wouldn't recommend this, it's too complex.
3. you have built in various rules that are likely to cause problems. For example, your selling at RS>75 is absurd, as you are then selling momentum in a momentum model. Why not just hold until the momentum breaks? Just as you don't get out at the bottom, you don't get out at the top. But that's momentum.
4. taking just the one top stock is suboptimal. Momentum has to unfold. But this way it constantly knocks you out of a value that was 1st last week and only 2nd this week. So you are always chasing the value that was best last week and end up underperforming. It is better for the models to take Top2 or Top3. Then the assets have time to breathe and you stay invested when an asset catches its breath or starts to rise.
Conclusion: Good idea, implementation still needs to be specified! 👍
But it also depends on the system. I had already experimented with a commodity GTAA model. But then I abandoned it for various reasons. I would have the following fundamental concerns with your model:
1. it is untested. Momentum models need parameters that exactly match the assets used, otherwise they won't perform.
2. momentum only works very moderately with commodities. Not all asset classes produce classic momentum. Oil, for example, does not work well, the movements are too erratic and driven by news. Momentum gains are more of a coincidence. You could check each commodity to see what the typical momentum intervals are and then use the average to determine the parameters. But I wouldn't recommend this, it's too complex.
3. you have built in various rules that are likely to cause problems. For example, your selling at RS>75 is absurd, as you are then selling momentum in a momentum model. Why not just hold until the momentum breaks? Just as you don't get out at the bottom, you don't get out at the top. But that's momentum.
4. taking just the one top stock is suboptimal. Momentum has to unfold. But this way it constantly knocks you out of a value that was 1st last week and only 2nd this week. So you are always chasing the value that was best last week and end up underperforming. It is better for the models to take Top2 or Top3. Then the assets have time to breathe and you stay invested when an asset catches its breath or starts to rise.
Conclusion: Good idea, implementation still needs to be specified! 👍
•
11
•
@Epi thank you for your well-founded answer and for the well-founded tips.
Takeaways for me:
1. check the universe
2. reconsider take profits
3. expand composition
4. definitely test
I'll start refining in the next few. Thanks ☺️
Takeaways for me:
1. check the universe
2. reconsider take profits
3. expand composition
4. definitely test
I'll start refining in the next few. Thanks ☺️
•
11
•