Investors need to do their part to mitigate risk:+1:
Diversification via holding multiple Darwins is your best bet for that, it drastically reduces the risk, and you get paid commission rebates for your wise actions.
Ah, understood…I was stuck at “why risk more than the trader”
Monte-Carlo simulation (daily and weekly return) do the job, except uncertainty calculations witch take into account inssuficient datas.
- Survivor bias : Sigma is computed with Monte-Carlo simulation without modeling the survivor bias. Futur probability is determined by the past so survivor bias product an under estimation on DD expectancy, especialy when you select good DScore.
- Rare events. A bad LA (as exemple but it is not the only cause), create rares events about DD. When it is about rare event, you need longer time and data (track record) to make an estimation. I suppose that this phenomenon is not modeled.
Thanks @CondorcetInvestment for this detailed answer
The problem is that to calculate a real accurate VaR you would need 3 years of trackrecord, the same as to factor out luck and state the edge of a strategy/trader.
The system of investable attributes is an attempt to find good trading without waiting 3-4 years .
Will reducing VaR to 5% help find good trading…on paper it should…
It would only slow down everything, the higher the risk the first you find bad traders and losing stuff.
Banks sell a lot of low risk crap, you understand it only after 10 years.
Time is much more important than money, you can’t recover it.
haha…sad but true…The moral of the story? Risk only what you can afford to lose, it will save you time! haha…
It is just the oposit. When you increase your VAR from 2% to 10%, you are REDUCING your Darwin risk by FIVE times (while the underlying strategy is risking more, so que equity of the trader, not the investors money that is linked to the Darwin).
I guess you missundertood my message. I wasn’t explaining how it works Darwinex right now, with a fixed VAR of 10%.
I was explaining what would happen if VAR in Darwins is not fixed at 10% (as it is right now), but they have a variable VAR with a maximum at 10%, meaning VAR is capped at 10% but could go even to 0,5%.
Here, I am not talking about underlying strategy but about Darwins.
I hope it makes sense right now what I explained in some previous message.
You wrote this:
And maybe I still don´t get it, but reading your example I understnd you are suggesting to an investor that investing 1000Eur at 10% is same than investing 5000Eu at 2%, and this is not correct. It is the oposit, if you invest in a Darwin 1000Eur at 2% is the same than investing 5000Eur at 10%. As lower the VAR is for a Darwin as high the risk it is for the investor. On the other side, it is the oposit for above 10% VAR. As higher the VAR over the 10%, the risk for the Darwin is lower (only the risk for the trader and his equity is higher). Please, let me know if I missed something.
BTW, I don´t pretend to be right, I am just trying to clarify this point (I think it is very important for both to understand it well, traders and investors).
I guess you still don’t understand what I want to say, so lets see if I can explain it a little better.
I wrote that sentence in response to this message:
@LenanaPeak was suggesting to have a maxium VAR of 10% if underlying strategy has a VAR above 10%, and have a variable VAR if underlying strategy has a VAR bellow 10%, which means that Darwins (not underlying strategy) would have a variable VAR, and not a fixed one at 10%.
Now, considering a variable VAR for DARWINS (forget about underlying strategy), it is similar to invest 1.000 EUR into a Darwin with VAR 10%, than invest 5.000 EUR into a Darwin with VAR 2% (VAR 2% in Darwins doesn’t exist right now, but it was what @LenanaPeak was proposing).
- 1.000 EUR * 10% = 100 eur maxium risk per month (statistically)
- 5.000 EUR * 2% = 100 eur maxium risk per month (statistically)
I hope I have explained this time a little better.
Does it makes sense now?
Yes, putting all together in this example, that´s correct and easier to undertand.
Just a comment: if you take the other example I pointed in my previous comment and you isolate off the whole context, it could lead to a mistake (as it happened to me), and because this is very important point I considered it worthed to make it as clear as possible. Also, because I have seen this mistake several times in the forum…
Thank your your patient.
Thanks! Probably my mistake for not being more specific.
Which VaR would you assess to this trackrecord?
I would state that monthly VAR at 95% confidence for this CTA is 5%
We have a max monthly loss of 8% and a max DD of 17%
The other way is classical economic theory equations.
Annual Sigma --> Monthly Sigma / MathSqrt(12)
95%/5% confidence --> Var=1.96 Sigma
Var(Monthly-95%) = 10.86 * 1.96 / 3.46 = 6.15%
(Annual Sigma = 10.86 from the previous link).
The most successful traders, work with VAR <5% (there are many examples, among which I follow, ZVQ, ERQ, THA, NTI, STV …) or a tad more, 6 or 7%.
There are very few good examples of strategies with risk of 10% or more.
But, if as is my case, we work with risk of the environment of 2.5% in the strategy (one of my darwins is HGO), the investor multiplies the movements of the same by 4 (in a not perfect but acceptable calculation). In addition, going from working with 2.5% to 5% to try to get closer to 10%, apart from that it is not at all clear that it is very desirable, it means changing lots and capital accordingly, and the SCORE manager of DWX takes a while more or less long in adapting to the new situation, so the SCORE, in that month, suffers …
But this I think is NOT the important thing: the fact is that for a VAR of 10%, we have an average annual target yield of 25% as very desirable, with a maximum loss of 20-25%, also annual. The famous ratio Return / maxDD in annual format, where anything greater than 1 is a great merit (because, after 5 years, if every year has occurred that case, the composite performance will be 1.25 ^ 5 = 3, ie , the 300% yield, while the maxDD will have stayed at that 20-25%.
Realistic goals and in fact, very very good.
But a 20-25% maximum annual DD is enough … the percentage deceives, because the investor sees what are the euros, dollars or pounds lost … perhaps it is little for a portfolio of € 1000, but much for a € 20,000 and very much for a € 100,000 or more … for me, go through -10,000 or -20,000 € losses seems to me excessive volatility (despite being very profitable in the long term!) And that if You get the returns, because if you had 50,000 pounds and after 5 years have multiplied by 3, and you have 150,000, that 20% (for the same VAR …) will be 30,000 pounds. If, after winning 100,000 pounds, a DD of € 30,000 is better, but the days go by and the number is undermining confidence …
Would not investors sleep much better, especially those who have more money and therefore more confidence, with a VAR of 5%?
And it is not a 12% of annual objective profitability an excellent result, that practically does not give you any other asset ??
What do the forum experts say?
According to this study by @JJENSLOPFAM the average is annualized 7% with 25% DD
This is my backtest: less than 14% leveraged, so 20% var
I suggested in Oct 2016 that Darwinex set a VaR floor for all strategies (back then the VaR of DARWIN was 20%):
In another post in Nov 2017, I talked about the idea of two Exchanges:
‘I think Darwinex should provide more options to attract traders who want to generate stable 10-15% annual return with around 5% drawdown, which institutional funds are very interested in. You cannot ask them to invest half of the money in DARWINs of 10% VaR. It is not like 2 * 1/2 =1. Darwinex Exchange could be divided into Exchange A and Exchange B. A different risk manager can be designed for DARWINs in Exchange B, not just 5% VaR.’