2Mo
Thank you of course for your work and ideas!
I have a few (critical) comments:
1.
The use of the TIPS indicator is of course obvious. I have already played through this. The mDD decreases significantly, which can be used. However, this increases the strategy cluster risk if you have 2xSpytips in your portfolio at the same time.
In addition, 3xGTAA is trimmed for high performance (incl. BTC) and with the TIPS indicator the model is often offline when certain assets rise.
My suggestion would be to optimize the 3xGTAATips model by selecting assets that perform positively with TipsNeg (e.g. gold).
2.
Your choice of leverage is strongly driven by the backtest, not by fundamental considerations. Therefore, there is a risk of overfitting. Behind the 3x leverage is the consideration of risk parity of all assets held based on BTC. This systematically reduces the risk, not only in the backtest.
I have a few (critical) comments:
1.
The use of the TIPS indicator is of course obvious. I have already played through this. The mDD decreases significantly, which can be used. However, this increases the strategy cluster risk if you have 2xSpytips in your portfolio at the same time.
In addition, 3xGTAA is trimmed for high performance (incl. BTC) and with the TIPS indicator the model is often offline when certain assets rise.
My suggestion would be to optimize the 3xGTAATips model by selecting assets that perform positively with TipsNeg (e.g. gold).
2.
Your choice of leverage is strongly driven by the backtest, not by fundamental considerations. Therefore, there is a risk of overfitting. Behind the 3x leverage is the consideration of risk parity of all assets held based on BTC. This systematically reduces the risk, not only in the backtest.
••
@Epi Your message seems to have been cut off.
1. that's right, MATT will also replace Spytips in my portfolio. I only use TIPS for the SPY, other assets are selected normally according to dual momentum. I have not yet divided the assets into negative and positive tips in general
2. this is also partly true. However, I have always used leverage 2 when there is one. With the exception of BTC because it is volatile enough and bonds because they are less volatile (EEM is technically also leverage 2 because cash is included). SPY is the other exception, but it is somewhat hedged by TIPS. Leverage 2 should also perform well
1. that's right, MATT will also replace Spytips in my portfolio. I only use TIPS for the SPY, other assets are selected normally according to dual momentum. I have not yet divided the assets into negative and positive tips in general
2. this is also partly true. However, I have always used leverage 2 when there is one. With the exception of BTC because it is volatile enough and bonds because they are less volatile (EEM is technically also leverage 2 because cash is included). SPY is the other exception, but it is somewhat hedged by TIPS. Leverage 2 should also perform well
•
11
•2Mo
@SemiGrowth Where is the message cut off?
1. tips also works very well for BTC. Logically, Tips is a liquidity indicator.
2. why actually leverage 2? Are there other reasons than optimized backtest?
1. tips also works very well for BTC. Logically, Tips is a liquidity indicator.
2. why actually leverage 2? Are there other reasons than optimized backtest?
••
@Epi the last sentence.
1. yes, for the QQQ TIPS doesn't work so well, I'll have to test it out
2. 2 is between 1 and 3 xD. The reason is that lever 2 reaches my target region of about 18-20% CAGR well. I don't need lever 3 in the width and lever 1 only delivers approx. 12- 15% CAGR
1. yes, for the QQQ TIPS doesn't work so well, I'll have to test it out
2. 2 is between 1 and 3 xD. The reason is that lever 2 reaches my target region of about 18-20% CAGR well. I don't need lever 3 in the width and lever 1 only delivers approx. 12- 15% CAGR
•
11
•2Mo
@SemiGrowth All truncated text ends at the last sentence. 😅
•
22
•
@SemiGrowth What is MATT?
••
@Redfox77 the strategy presented in the article
•
11
•@SemiGrowth I should finally get myself a reading library...
•
22
•