New Leverage stat messing up right now

The new Leverage stat is messing up right now, at least for my stock strategies. Example:

That’s currently showing a range of 2.06 - 3.04 leverage… but it’s an IRA account at IB that’s driving it via Broker Platform, so it’s actually never been over 1.0 leverage, by law.

You’re trading a lot of TQQQ, which is a 3x levered ETF. The way C2’s leverage is calculated, it will consider all sources of leverage, including both leverage that is inherent to the instrument and leverage where you’re actually borrowing money from the broker. This is done so that the leverage stat is standardized and can be compared across different markets (stocks/futures/forex/options). C2’s leverage stat is in this sense a representation of risk. A dollar invested in TQQQ carries 3x the risk of a dollar invested in QQQ.

Ah, thanks. I was pretty sure that when it was first introduced, my strategies all properly showed the actual amount of leverage as considered by brokerages (ie, i’m not using margin). But now it’s different. It is “Leverage”, and not “Margin”; my mistake.

Glad I could help.

Leverage in this case is not just use of margin, but also how much leverage there is inherently in the instrument you’re trading, as compared to its underlying benchmark.

Has to be done this way so that the stat makes sense across all instrument types and is directly comparable between, say, a system that trades futures and a system that trades stocks, as a measure of how much $ exposure there is to the underlying market’s price changes.

1 Like

You are using Max 3.04 leverage only but DD 40%. That’s mean “Higher leverage = greater risk of system” is not correct. In other word leverage is nothing to do with the risk of a strategy.

"Collective2 calculates the maximum leverage used by a strategy in each day. We then display the average of these measurements (i.e. the average daily maximum leverage) and the greatest of these measurements (maximum daily leverage).

Leverage is the ratio of total nominal value controlled by a strategy divided by its Model Account equity. Generally higher leverage implies greater risk.

Example of calculation:
The Strategy buys 100 shares of stock at $12 per share.
The Model Account equity during that day is $5,000.
The leverage is: $1200 / $5,000 = 0.24

This is a useful measurement, but it should be considered in context. This measurement doesn’t take into account important factors, such as when multiple positions are held that are inversely correlated. Nor does the measurement take into account the volatility of the instruments being held.

In addition, certain asset classes are inherently more leveraged than others. For example, futures contracts are highly leveraged. Forex positions are often even more leveraged than futures."

You are using Max 3.04 leverage only but DD 40%. That’s mean “Higher leverage = greater risk of system” is not correct. In other word leverage is nothing to do with the risk of a strategy.

All other things equal, a strategy with higher leverage will be riskier than a strategy with lower leverage. If his strategy was running at 1x leverage (let’s say QQQ instead of TQQQ), and all other factors were kept the same, the max DD would have been ~13%. If it was 6x leverage, the DD would have been ~80%.


How do you know 2 strategies have all other things equal or not?

Strategy A with higher leverage but lower DD and higher return.

Strategy B with lower leverage but higher DD and lower return.

Which one would investors like?

Here, many people like strategies with 30-50% profit per month and 100% drawdown. Such a truth of life. Lose. Get the pain. And again for the new. Sadomasochists some.

That’s a nice strawman you’ve set up there.

Like I said, all other things equal (use of stops, adding to losers, quality of entry signal, etc.), more leverage = higher risk of failure.

p.s. This isn’t about what investors like, but about what indicates higher chances of failure, empirically.

1 Like

The leverage calculation as explained seems to be a poor proxy for risk in the case of options since the delta of the options is not taken into consideration.

For example, a portfolio that invested 100% of the initial assets in treasuries and then exclusively invested the interest earned into buying low cost deep OTM options could achieve a very high leverage score since each option contract purchased would disproportionately add to the nominal value of the portfolio. In reality, this portfolio would be about as risky as sitting in cash and provides a simple counter example to the accuracy of the leverage calculation as a universal proxy for risk.

I like what C2 is trying to do with the leverage calculation. However, I would like it more if they would attempt to address some of the gaps that the current implementation appears to create.

If options are not handled in the way I have described, please help me understand how the option delta is factored into the calculations?

I see Dwight already highlighted my fundamental concern with the current implementation.

Just to be clear I think it is a fantastic and very useful too. I also think that including leveraged ETFs in the calculations is smart. I was mostly trying to make the point to people that did not think highly of it that they need to lower their expectations of what it can do. It is a great tool that I use a lot when looking at strategies now.

Under the current implementation, options “leverage” calculations assume a delta of 0.5, which is not always perfect, but which, for most cases, seems a pretty reasonable assumption (i.e. at the money, with time left). I think implementing a much more detailed model than this would not be worth the computational overhead.

And sure, for those cases of far OTM options, or options near expiry, then these leverage calcs will have only limited value.

Mathew, thanks for the additional details on the current implementation. If it proved technically feasible to vary the delta values, I do believe it would enhance the value of the metric for portfolios holding options.

I also understand, you probably have a long list of enhancements in your backlog.