6Mon
Thanks for the great contributions, but I still have the following questions:
- In step five, the risk is distributed to the "top 2 or 3". What exactly does that mean? What criteria are used to weight them? And shouldn't this worsen the overall performance? After all, these two assets have a worse performance, plus you make more trades
- Both the wikifolio and your updates only include BTC and gold as assets alongside the Nasdaq100. Did you just leave out the remaining assets to maximize the return for the 3x leveraged portfolio or is there a deeper reason for this?
-> What basket of assets are you running with your main GTAA portfolio, if you have one? Of course I want to research and backtest it myself, but can't find any tutorials on TAA at Portfolio Visualizer (never get the same results as here in the post)
- In step five, the risk is distributed to the "top 2 or 3". What exactly does that mean? What criteria are used to weight them? And shouldn't this worsen the overall performance? After all, these two assets have a worse performance, plus you make more trades
- Both the wikifolio and your updates only include BTC and gold as assets alongside the Nasdaq100. Did you just leave out the remaining assets to maximize the return for the 3x leveraged portfolio or is there a deeper reason for this?
-> What basket of assets are you running with your main GTAA portfolio, if you have one? Of course I want to research and backtest it myself, but can't find any tutorials on TAA at Portfolio Visualizer (never get the same results as here in the post)
•
11
•
6Mon
@Simon_n You've dug out an old classic from my days of busy model making. Very good! 😅
To your questions:
- "Top 2 or 3" means that you take the 2 or 3 assets with the highest momentum sum of the defined momentum series into the portfolio. The weighting can be linearly increasing, decreasing or equal. I prefer equal weighting because it significantly reduces portfolio turnover. The overall performance does not suffer because the negative influence of mean reversion diminishes after 1 month. In other words, the assets have some time to develop and do not always have to have been at the top of the momentum in the last month. This also reduces portfolio turnover.
- The Wikifolio and my portfolio only ever contain the current allocation of the top 3 Mom stocks. The others are in the queue.
- I now only run 2 GTAA models, the 3xGTAA and a (defensive) unleveraged model via my pRV shell. With the latter, I have tried to create a reasonably high-performance model from the 200 ETFs available. Unfortunately, mylife doesn't have everything, e.g. no gold ETFs (I have to use gold mining stock ETFs).
- The fact that you never get the same results as in my backtests is probably because you have used other momentum series and other time periods. The exact figures don't really matter. The main thing is that the direction is right: constant returns with significantly reduced risk.
To your questions:
- "Top 2 or 3" means that you take the 2 or 3 assets with the highest momentum sum of the defined momentum series into the portfolio. The weighting can be linearly increasing, decreasing or equal. I prefer equal weighting because it significantly reduces portfolio turnover. The overall performance does not suffer because the negative influence of mean reversion diminishes after 1 month. In other words, the assets have some time to develop and do not always have to have been at the top of the momentum in the last month. This also reduces portfolio turnover.
- The Wikifolio and my portfolio only ever contain the current allocation of the top 3 Mom stocks. The others are in the queue.
- I now only run 2 GTAA models, the 3xGTAA and a (defensive) unleveraged model via my pRV shell. With the latter, I have tried to create a reasonably high-performance model from the 200 ETFs available. Unfortunately, mylife doesn't have everything, e.g. no gold ETFs (I have to use gold mining stock ETFs).
- The fact that you never get the same results as in my backtests is probably because you have used other momentum series and other time periods. The exact figures don't really matter. The main thing is that the direction is right: constant returns with significantly reduced risk.
••
6Mon
@Epi Yes, I plan to try out either a SPYTIPS or GTAA variant myself, and to make sure it doesn't go wrong I need/want to have all the knowledge I can (even if my fingers hurt after scrolling back through the posts)
The fact that the other assets that didn't make it into the top 3 are simply not visible because they are in the queue makes sense, of course.
But it is also interesting that the other, initially weaker assets in the top 3 can still develop and thus the return does not suffer.
I'm also trying to find the "perfect" variant, including cooldowns, SMAs etc., as another user is also trying to do for SPYTIPS. (Do you also use a GD in your variant, or have you found another method to reduce strong drawbacks due to unfavorable timing?)
But I'm afraid that this won't work with Portfolio Visualizer alone, Pinescript in TradingView quickly becomes expensive. The alternative is probably Python. Did you do all your backtests with Portfolio Visualizer only? And why WisdomTree instead of Euwax II? Sorry for all the questions 😁
The fact that the other assets that didn't make it into the top 3 are simply not visible because they are in the queue makes sense, of course.
But it is also interesting that the other, initially weaker assets in the top 3 can still develop and thus the return does not suffer.
I'm also trying to find the "perfect" variant, including cooldowns, SMAs etc., as another user is also trying to do for SPYTIPS. (Do you also use a GD in your variant, or have you found another method to reduce strong drawbacks due to unfavorable timing?)
But I'm afraid that this won't work with Portfolio Visualizer alone, Pinescript in TradingView quickly becomes expensive. The alternative is probably Python. Did you do all your backtests with Portfolio Visualizer only? And why WisdomTree instead of Euwax II? Sorry for all the questions 😁
•
11
•6Mon
@Simon_n No problem.
It might be a good idea to set up a website with all my posts in an orderly fashion and put a link to it.🤷
A fundamental problem with momentum models is overfitting and the phenomenon that models are regularly and systematically found at their peak performance. That's why I always try to keep things as simple and robust as possible.
The cooldown variant of Spytips is already cool 😏, but I still find it a bit too complicated. Systematically, however, I have no objections.
I try to avoid large drawdowns not only through model optimization, but also through strategy diversification. It worked well in March.
If necessary, a good AI can also build you a backtest bot. You just have to know what you want and tell the AI.
By Wisdomtree you mean the 3x Lev. gold ETC? Well, EuwaxII is unleveraged. Therefore.
It might be a good idea to set up a website with all my posts in an orderly fashion and put a link to it.🤷
A fundamental problem with momentum models is overfitting and the phenomenon that models are regularly and systematically found at their peak performance. That's why I always try to keep things as simple and robust as possible.
The cooldown variant of Spytips is already cool 😏, but I still find it a bit too complicated. Systematically, however, I have no objections.
I try to avoid large drawdowns not only through model optimization, but also through strategy diversification. It worked well in March.
If necessary, a good AI can also build you a backtest bot. You just have to know what you want and tell the AI.
By Wisdomtree you mean the 3x Lev. gold ETC? Well, EuwaxII is unleveraged. Therefore.
••
6Mon
@Epi All right, thanks for the answers! Keep following the updates with interest
•
11
•