Exposure of systems?

How do we determine the exposure (i.e. % of capital in market over the time-frame) of the systems on Collective2?



This is needed to compare across the various systems.

- Annual Return isn’t sufficient as it doesn’t indicate what % of the capital was utilized.

- Also, can’t use Trade Duration X # of trades, because it doesn’t indicate the average position size.



This is a general question pertaining to all systems on Collective2 - need to quickly evaluate this metric across systems.



thanks,

Kiran

[LINKSYSTEM_57830847]

I was looking for the exposure time in market for C2 systems. Is there a way to find this out? If not, please consider this as a feature request

(I may be biased here due to my systems, which at the best has 50% exposure time (NQ Positional) and otherwise exposure on my systems is about 10% i.e most of the times systems are idle and hence lesser risk )

I think it is a valuable parameter for subscribers to choose a system

Doesn’t the average leverage calculation measure this?

probably in a way that considers time of exposure as well but in terms of leverage. I was thinking it terms of Exposure % as Market exposure of the trading system calculated on bar by bar basis. Sum of bar exposures divided by number of bars. Single bar exposure being the value of open positions divided by portfolio equity.

I would say that the average leverage calculation does the same thing. If a strategy buys TLT with 10% of the portfolio and leaves the rest in cash and that the average leverage calculation would read 0.1 assuming it maintains that 10% allocations buy buying or selling off as needed. If it is just one buy or sell it would then move up or down from 0.1 as TLT changed.

Let me rephrase it - lets say two systems - S1 and S2. Lets also forget about leverage. Lets say both of them buy/sell QQQ when there is a signal.

If S1 buys QQQ with his total equity at the beginning of the year and sells at the end of the year and makes 10% at the end of year, his return is 10% and average leverage is 1 and exposure% is 100%.

If S2 buys QQQ with his total equity at the beginning of the year and sells at end of January making 10% and trades no more for the year , his return is 10% for the year, average leverage is still 1 , and his exposure is only 1/12th of S1.

i.e S2 is a better system.

I don’t think that the average leverage in S2 would be 1, the way C2 calculates it. I think it would be 0.083.

if you look at the system

average leverage is approximately 3, and if you click the leverage to overlay on the chart, you will see system leverage when in trade is 3 (it buys/sells tqqq) and system has idle times when it is in cash. If calculation is according to your definition, I believe the average leverage will be definitely under 3.

You are correct. That does not seem to fit with their definition. Perhaps @MatthewKlein could clarify what we are not understanding.

2 Likes

Maybe they only count days when you have a position and disregard zero days? I was eyeballing my strategy as well and I don’t think my average leverage is 0.41 over the history of the strategy.

yup. I think so. I think the zero days also should be considered, either in the same statistic or as a different statistic of exposure.

2 Likes

Of course. Zero risk those days.

1 Like

@EvaRossmann and @GardCap interesting findings. I had never noticed that zero days seem to be neglected. I certainly agree they should be factored in.

3 Likes

I think calculation should include not just in terms of days, but at a more granular level. A system that exposes itself to only 1 hour per day should not be treated in the same manner as another strategy that has 8 hours exposure. (the former has only 1/8th exposure, assuming same leverage employed on both)

3 Likes

any feedback on this @MatthewKlein ? Since this parameter is so important for leaderboard ( I have noticed a cutoff of 3.5 on the average leverage (how C2 calculates as of today) has a big impact on system ranking). Even otherwise , for system selection, I believe it is an important criteria