Make sense… this DD in the nasdaq is not happening quite often… anyway I made the rules and so I will follow them to see if the porfolio can produce gain in a well defined time window (6 month- 1 year).
Sold half position of TKT new historical DD (>-7%)
are not we all…
Up to my kowledge SYO is algorithmic trading, so there is no “insistence”
You design and test a ruleset and you execute it, period.
No visions, opinions…
I have made some analysis and consideration of the Portfolio.
Let’s consider only october: 13 out of 20 darwins have registerd a loss.
LA high: -2.39% 5/10
LA low: -1.12% 7/10
Wrost LA low:
Wrost LA high:
this was the wrost month since 2018 of this portfolio.
I’m not very surprised to see that using Dscore as a selection I will have those results: Performance weight very much in Dscore, but performance is quite often mean reverting, so if we consider a single darwin the rule not to enter at top can be not right, but when we consider a portfoglio of many darwins, not to enter in a top can be a good rule.
On the other side the Low La score group performed lower than the High La score group: I think that in Low La score group the impact of performance can be more pronunced that in the other group. In fact a Low La score means that when the darwin loses it may loose very fast… so a low la darwin can have many periods of good performance and then suddenly a “big” drop. High La darwin tend to have prolunged DD and less steepy. So entering in a “peak” of performance (i.e. high Dscore) can be more dangerous for low LA score.
I’ve decided to change the way I wanted to follow this demo portofolio: since its creation darwinex made available API to beta tester and I’m one of them. With the API it can be possible to test rules and construct portfolio in a better (I hope ) way. Also monitoring a portfolio will be easier. So I will continue to follow this portfolio but with only one rule: if a darwin fall below 55 at the end of the month I will close all the position in that darwin. I will update so every two weeks to see the performance update.
Today the porfoglio is still in DD:
-2.31% in November.
Only six Darwin are making profit since the beginning:
TGB 4.95% KVL 3.88% , BSL 3.4% , STV 2.16% , THA 1.83% , CJN 0.98%
UEI -11.72%, ZTY -10.85% LZL -9.75% ABH -8.52% VRT -10.05%, TKT -9.33%
4 DARWIN OUT of 6 of the best performer are in the LA high score
5 DARWIN OUT of 6 of the wrost performer are in the LA low score
For now the diversification in two groups is not giving a “boost”.
I think we won’t be able to define a better ruleset until some months. I mean, once the current market conditions will be over, we’ll be able to choose some darwins that performed nicely these days to hedge those good darwins that despite of their good track record, performed horribly under these conditions.
-0.63% In november.
The portfolio is currently reducing the DD.
I’m more interested in the second image: some darwins are getting a positive result thanks to the Huge positive divergence… are those Numbers real? I mean that this is a demo so the divergence Factor may be over/lower estimated…
I think divergence is different from demo to live.
I have DCD on live with a divergence of -2.28% (stacked since february 2018)
On the opposite side I have PUZ at +2.28% (stacked since July when i sold and rebought it to rebalance)
+0.44% in December after a -0.63% in November and a flat start for the year (considering the YEN related flash crash, I think that this can be considered a good start).
Diversification between the two groups maybe is starting to work out. Now I want to see some good profit.
Have you sold also half of your position on KVL ? (highest historical drawdown)
No, rules were changed: Test-Demo Portfolio LA low score and LA high score now I only close a darwin if it goes below 55 at the end of the month.
It makes sense, in this way you often bought high but at least you are not selling low.
As I have written I was aware that the selection made was biased and not rightly timed, but my point here was to test a selection that besides the near term correction that was possible, could prove to be profitable at least in 6-12 months. That’s my goal to prove that it’s important to:
- do not have Darwin with a constant DScore below 55 (I was thinking about 50, but if I take the score at the end of the month this can avert to discard a darwin only for a few points)
- diversify through LA score (it would be useful to identify other typology…)
- to not use a stop loss but using the portfolio to reduce DD.
Do you still think that 5 is the right threshold to divide trendfolowers and contrarians?
I am using 6 and maybe it shoud be increased further.
I decided for 6 because my KVL is a contrarian and LA floats from 4.5 to 6.
LA is useful to cathegorize the style but we have to consider that is has been designed to rate “good” and “bad” trading.
Probably your way is better. The point is to identify different types of trading.
Exact and it is working!
I think your is a very good approach, just suggesting a mild improvement.
I think “bad trading” is something like LA < 1.5 so that could be considered like a baseline.