Thanks for all the detailed information. Of note, I think it’s very important to inform subscribers that the system is based on buying the ETFs on the first day of the month— because it makes a difference. I bought XLE on June 2 and that was quite different than buying on June 1. I think I asked about this in May before subscribing, and you recommended buying at least in the first day or two. It’s important to Inform subscribers that the system is based on purchasing the ETFs on the first day of the month and purchased by the end of the day-if that is how the system is set up. Otherwise, people may get very different results. Perhaps it all comes out in the wash and that’s something you could also address—as sometimes their purchase price could be better, rather than worse. And, as markets fluctuate during the day, perhaps it would be helpful to know exactly how you calculated back testing by indicating the timing of the etf purchase.
Thank you, yes, that’s very helpful. I think it’s meaningful to inform subscribers that you buy on the first day. As although it may come out in the wash in the long-term, it makes it more predictable to know that you are purchasing the first day and that’s how your numbers are calculated. I appreciate you constantly answering my questions as I’m trying to determine if this is a fit for me. And perhaps this is something you already have laid out clearly in your description of the service. And I do understand there’s no way to determine exactly how people will do individually and that their numbers may vary, but knowing that your numbers are based on buying it the first day, for me, it’s very important-it might not be for other people.
It's important to understand that two people following the same strategy will never see the exact same returns. Prices can move up and down a lot from day to day and also within a trading day. This will sometimes come out in your favor, sometimes not.
You can't conclude that buying on the second trading day is always worse than buying on the first, just as you won't know if buying early in the day is better than late.
What the system tells you is the price-based probability that certain assets will perform well over the coming month. The earlier I buy after the signal, the better the chance is that the price hasn't moved a lot already.
The numbers in the algorithm are based on the monthly closing prices, and I always buy on the first trading day. That is the intention. Sometimes I get a better price than when the signal came, sometimes it's worse. In the long run, these differences tend to balance out each other.
I made a replication study of this approach, which obviously isn't a 1-to-1 match due to the proprietary algorithm, but I still get very similar performance and the same drawdown behavior. I also ran a fairly extensive rebalancing timing luck experiment, and at least in my case, there is indeed a huge amount of CAGR dependent on mere luck. If you rebalance around, say, the 15th of the month, you get drastically worse returns while following an otherwise identical strategy. That isn't alpha. Unfortunately, it's rebalancing timing luck.
Thanks for the input, I appreciate your work in studying the strategy! And as you mention yourself, of course it can't truly be evaluated without knowing the full proprietary algorithm.
The finding in your study is interesting, and RTL is one among many factors in backtesting that are related to luck. In my opinion, one should see a backtest as the upper interval of the forward expectation, while the OOS data (live trading) is what confirms or disproves the robustness.
The choice of month-end rebalancing was to keep things simple and to avoid the potential curve fitting that would occur if you searched for the optimal rebalancing time in every month.
The EUR-version of the strategy had already outperformed the MSCI World Index in 20 out of 21 years when I started trading it. In the following 5 years, where I have traded it with real money, it has continued to outperform every year. The probability of outperforming in at least 25 out of 26 years by pure luck would be 0.0000402%. Given the simplicity of the proprietary algorithm, this multi-decade performance would be extremely unlikely to be driven by pure luck.
This doesn't mean we can necessarily expect the same level of outperformance in the future. In financial markets you can never guarantee anything.
Thanks for all the detailed information. Of note, I think it’s very important to inform subscribers that the system is based on buying the ETFs on the first day of the month— because it makes a difference. I bought XLE on June 2 and that was quite different than buying on June 1. I think I asked about this in May before subscribing, and you recommended buying at least in the first day or two. It’s important to Inform subscribers that the system is based on purchasing the ETFs on the first day of the month and purchased by the end of the day-if that is how the system is set up. Otherwise, people may get very different results. Perhaps it all comes out in the wash and that’s something you could also address—as sometimes their purchase price could be better, rather than worse. And, as markets fluctuate during the day, perhaps it would be helpful to know exactly how you calculated back testing by indicating the timing of the etf purchase.
Thank you, yes, that’s very helpful. I think it’s meaningful to inform subscribers that you buy on the first day. As although it may come out in the wash in the long-term, it makes it more predictable to know that you are purchasing the first day and that’s how your numbers are calculated. I appreciate you constantly answering my questions as I’m trying to determine if this is a fit for me. And perhaps this is something you already have laid out clearly in your description of the service. And I do understand there’s no way to determine exactly how people will do individually and that their numbers may vary, but knowing that your numbers are based on buying it the first day, for me, it’s very important-it might not be for other people.
Hi Lindsay
It's important to understand that two people following the same strategy will never see the exact same returns. Prices can move up and down a lot from day to day and also within a trading day. This will sometimes come out in your favor, sometimes not.
You can't conclude that buying on the second trading day is always worse than buying on the first, just as you won't know if buying early in the day is better than late.
What the system tells you is the price-based probability that certain assets will perform well over the coming month. The earlier I buy after the signal, the better the chance is that the price hasn't moved a lot already.
The numbers in the algorithm are based on the monthly closing prices, and I always buy on the first trading day. That is the intention. Sometimes I get a better price than when the signal came, sometimes it's worse. In the long run, these differences tend to balance out each other.
I made a replication study of this approach, which obviously isn't a 1-to-1 match due to the proprietary algorithm, but I still get very similar performance and the same drawdown behavior. I also ran a fairly extensive rebalancing timing luck experiment, and at least in my case, there is indeed a huge amount of CAGR dependent on mere luck. If you rebalance around, say, the 15th of the month, you get drastically worse returns while following an otherwise identical strategy. That isn't alpha. Unfortunately, it's rebalancing timing luck.
Thanks for the input, I appreciate your work in studying the strategy! And as you mention yourself, of course it can't truly be evaluated without knowing the full proprietary algorithm.
The finding in your study is interesting, and RTL is one among many factors in backtesting that are related to luck. In my opinion, one should see a backtest as the upper interval of the forward expectation, while the OOS data (live trading) is what confirms or disproves the robustness.
The choice of month-end rebalancing was to keep things simple and to avoid the potential curve fitting that would occur if you searched for the optimal rebalancing time in every month.
The EUR-version of the strategy had already outperformed the MSCI World Index in 20 out of 21 years when I started trading it. In the following 5 years, where I have traded it with real money, it has continued to outperform every year. The probability of outperforming in at least 25 out of 26 years by pure luck would be 0.0000402%. Given the simplicity of the proprietary algorithm, this multi-decade performance would be extremely unlikely to be driven by pure luck.
This doesn't mean we can necessarily expect the same level of outperformance in the future. In financial markets you can never guarantee anything.