I'dlike to share with you my research now I have some hindsight on the model.
This portfolio is running for 5 full months now, and have paid its Q1 perf fees.
The "spirit" behind this portfolio is a mix of maths, and simply... common sense (assomptions, deductions, observations), but operating under a predefined systematic model for trade execution.
The goal of this portfolio was:
to prove that correlation is more powerful to fix / stabilize a model than individual performance, or % allocation, if we assume that all assets have an equal statistical risk (wich is a topic that could be discussed later on).
to prove (me and) investors that focusing all our efforts on choice is the wrong way, and is useless because
1) it leads to destructive behaviors (ie, jumping from a DARWIN to another, buying a DARWIN @136, then sell it @135, the buy it again @141 then sell it @139, etc...). This behavior is in all point comparable to the trader always searching for the Graal... unable to stick to a process and so, systematically losing. And
2) Focusing on the quality of diversification on non-correlated assets, watching the portfolio as global, and not each DARWIN as an individual, makes you able to stick to your model, and the diversification compensates for the risk. For this reason, I've added randomness in its composition, you'll see on the chapter below.
to establish a simple and scalable process, that anyone can apply very easily in 5 steps, answering 5 questions.
To be honest, this kind of results was totally unexpected at the start of the experiment. The robustness, the volatility vs return, the overall behavior, are far above all my initial forecasts. For this reason, I'm taking some risk for the future posting this... why? because for the time being, it is robust and profitable!
So... enjoy, it is free lunch guys!
1. What are my choices?
-50% based on filter [D-Score >65] + [Drawdown -15% last year] (D-Score criteria could now be upgraded to >70)
-50% based on discretionary choices with only one criteria: they have to be non-correlated with the above, and accross them.
2. How many DARWINs?
12 to 14 DARWINs
why? Because I want to optimize my NET expected return (risk vs. diversification rebates) according to this chart
Equally split in all DARWINs
Just like @DarwinexLabs -> http://blog.darwinex.com/darwin-filters-alternative-markowitz-portfolio-theory/ , I think that a more or less equal allocation on quality DARWINs will produce a more robust /stable result than a strict MPT model, in the Darwinex environment.
4. Investment Timeframe?
I don't want a lot of rotation in my portfolio because:
- how could you do some stats and stabilize your overall performance if you jump from one DARWIN to another every day/week/month?
- I want to put time in my favor, for a recovery in case I enter on a DARWIN in a losing streak
- diversification is working for me and my global risk
I dont want DARWINs that are correlated more than +0.6 in porftolio
If it happens for more than 1 full month, one or the other must be replaced
Operational methodology & time management
- 50% cash / 50% DARWINs (or use 2:1 leverage to do so)
- Review of assets in portfolio only once a year
- Rebalance quarterly (cash in profits/losses and divergence, restart of equal allocation)
- Monthly arbitrage
If a DARWIN loses 7% or more in one month or less, I bought 1 unit more (if my initial investment is 200, 200 is a unit) to play the recovery (that's why we are 50% cash and 50% DARWINs at the start).
The speed of the loss is more important than the size here, we are searching for a "statistical accident". A DARWIN that loses 1.5% per month, 5 successive months for example, does not meet the criteria.
When do I sell my arbitrage:
- If the DARWIN recovers half of the loss between the peak and my price on this trade
- One month after, regardless of any other aspect (winning or losing) because I don't want to be over invested too much time on a DARWIN that is in a losing streak on a bigger scale (long term cycle).
based on 10.000$ (using leverage 2:1)
To switch from DEMO to LIVE environment, I've added a rule of entry.
So I won't jump in and buy all selected DARWINs, but will apply my arbitrage rule to enter my initial investment on each individual DARWINs, until my portfolio is complete.
Additionnally, because JMC is full now, I had to replace it. DNA was disappointing, but maybe I will keep it anyway to stick to the initial model in demo (thinking about it...).
Here's the list for the LIVE Portfolio
Actually in portfolio:
$PIK - https://www.darwinex.com/darwin/PIK.4.23
$NTI - https://www.darwinex.com/darwin/NTI.4.12
$DWC - https://www.darwinex.com/darwin/DWC.4.20
$PLF - https://www.darwinex.com/darwin/PLF.4.1
$EZX - https://www.darwinex.com/darwin/EZX.4.5
Waiting for a drawdown to buy:
$ERQ - https://www.darwinex.com/darwin/NTI.4.12
$LVS - https://www.darwinex.com/darwin/LVS.4.20
$DLF - https://www.darwinex.com/darwin/DLF.4.7
$UAF - https://www.darwinex.com/darwin/UAF.4.13
$KWC - https://www.darwinex.com/darwin/KWC.4.19
$VFL - https://www.darwinex.com/darwin/VFL.4.1
$THA - https://www.darwinex.com/darwin/THA.4.12
$PUL - https://www.darwinex.com/darwin/PUL.4.6
($DNA - https://www.darwinex.com/darwin/DNA.4.4)
Since this portfolio surprised me very positively in term of results, all the goals discussed earlier have been reached.
As a natural process, the goal is now upgraded Let's run it in REAL and see the (hopefully profitable) results in 6 months guys!