Strategy risk not apparent based on readily available data

I was wondering what do people think of the risk inherent in what they can “see” vs. what they “cannot see” on the C2 platform (see = what is readily apparent, cannot see = something that is there but requires a lot more digging to be revealed)

As we all know that strategy managers can create multiple accounts and/or take strategies private if they want. I tried to get thoughts from C2 on this and they sent me their philosophy. The gist of it is that -

  1. “At C2 we believe that the “person” behind a trading strategy may not be important. The only thing that matters is the track record of the strategy.”

  2. “That’s why we moved in a different direction, and try to provide tools which may already be familiar to you that we hope can help investors make informed decisions.”

Fair enough, I thought. So I tried to evaluate the track record of the strategies that I could “see” i.e. the ones that were public at the time. I came across 2 strategies from the same manager that looked good (the first one looking way better than the second) -

  1. SteadyReturns
  2. Alpha1st

I subscribed to both thinking -

  1. Risk/reward looks palatable.
  2. Even if there is a drawdown, I can take it and keep going and eventually it will recover.

So far, so good.

Then, I came across some forum posts that drew my attention to the fact that there is a way to get the trade CSV of strategies that have been made private. So I decided to do a bit of digging on the manager that I had subscribed to. He has created 10 strategies, out of which only 3 are public.
I downloaded the CSV files for all the 10 strategies. Since C2 has a handy P&L column calculated already, it was easy to plot the P&L chart for these (these are P&L charts, not equity charts. You can add starting equity to it to get the equity chart). I am attaching the P&L plot of 2 of this strategy manager’s strategies.

LXC201
ZLXCkilled4NoStopAlgo

I realized something was wrong. Some of the strategies that were made private were going up in a straight line and then had a massive drawdown. So I decided to look under the hood i.e. look in the CSV files to see what trades caused it. I realized a few things –

  1. The holding periods of the catastrophic trades was much bigger than the average. Even though the average holding period of the manager is less than 1 day, when things start to go wrong, they could hold a trade for up to 1 week. Essentially, they resort to hoping instead of taking the loss

  2. In some cases, they doubled, tripled, quadrupled the risk. When they ran out of MES contracts to trade, they resorted to using the ES contracts (something that is 10x of MES) to amp up the position to make back the loss

  3. When all else failed and things didn’t revert, they “killed” the strategy and made it private. This makes discovering this information in the future an order of magnitude harder as new users (even some old users) will not know how to retrieve CSV files for private strategies.

This got me thinking. Evaluating the public strategy is not telling me the big risks lurking underneath -

  1. When I thought the risk/reward is palatable, it actually is not. I have no way of controlling this trader’s behaviour in bad times. Since they don’t trade their own money, they could resort to cowboy behaviour to make it all back as the subscription revenue from a strategy in drawdown is as good as not having the strategy

  2. When I thought that I could recoup a potential drawdown from their public strategy, I was wrong. In the past, after amping up the risk and still being in a drawdown, they just shut things down and move on.

Essentially, I see the nice equity chart, subscribe, make some money for a bit and then lose a ton of money before the strategy shuts down and goes private. Am I wrong in thinking this?

I understand the platform has an economic incentive to have people subscribing to strategies and this is an inconvenient detail that is better hidden under the carpet by making strategies private. But should we not be doing more to make this information more readily apparent?

3 Likes

there is enough data in the public information of many of similar strategies - the win ratio and the leverage spikes.

1 Like

Information that these two strategies are dangerous is already available - almost 90% win rate for both of them. Having avg winner twice less than avg loser adds another reason to skip them.

1 Like

It’s funny, I have both of those strategies on my watch list. I have been amazed at how they have kept doing well for so long but eventually, they are likely to have the same issue you have uncovered in your research (good job!). I am waiting to see if/when they blow up since they employ the Martingale technique. Martingale System: What It Is and How It Works in Investing, which is basically doubling down after a loss, works until the strategy runs out of money.

This is something I always check when I see that kind of equity curve and a low loss rate. In the Trading Record, look for sudden large increases in the total number of contracts traded (Qty column). Click on Show Autotrade Data, and look at the actual trades. For example, in Alpha1st, look at the trade on 3/9/2023 where the number of contracts suddenly jumped to 35. He only traded one or two contracts at a time but the total exposure was 35 contracts before he won the battle. Also be aware that this system has been rescaled downward twice which makes the Qty column number on earlier trades smaller so you have to open the trades to see the actual exposure at the time.

Remembering that the stock market is not there to make us money, it’s there to take all the money we have, a Martingale technique can work for some time, as these systems demonstrate. They may not have worse outcomes in the long run than strategies that do not employ this technique. As with all strategies, you need to be prepared to get off the ride. Stops may save you but will likely give you worse performance.

The hardest thing with these types of systems is that when they go down, it’s fast and often for the last time. But what a ride until they do. I’m not that kind of adrenaline junkie.

2 Likes

I agree just have your own overall stop for the system , treat systems like stocks and have your own stop .

1 Like

The average annual return of the US stock market is about 10% a year (for almost a century now), dividends included, so obviously it does not “take all the money we have”, unless you meant something else.

Not necessarily, because there are also plenty of losing systems with 3 to 1 reward/risk ratio.

That part was meant to be read by people with a sense of humor.

1 Like

That is the answer to alot of issues here.

One trade leader has created a half dozen profiles over the last 5-6 years, blown up systems every time and gone private only to emerge with a new profile/system a week later. However he has nice looking equity curves, pays C2 to list them and gets new subscribers every time. C2 makes money off of him and his subscribers and he makes money.

But if subscribers actually made money on systems like this, this site would have far more interest than it does currently.

But the red flags are there in his trading style in his old systems and the new ones . Plus as we said subscribers really should have a system stop level even if the developer has a good reputation and a great past trading history.

2 Likes

Absolutely, there are plenty of trading systems that look great and “safe” at the beginning.
And then they blow up fast, sometimes years later, without warning.

That’s why C2 investors must protect their account at all times, regardless of the track record of any particular strategy.

1 Like

There are a lot of different losing systems. But imho trader who wins more and loses less in average doesn’t need to have high skills in market direction guessing, compared to the trader which in average loses more and wins less. Latter one needs to use tricks like martingale or waiting for the trade be positive. Unless he has excellent market direction guessing skills.

1 Like

Why not put a “community risk warning” on managers and/or strategies that are prone to bad behaviour?

For example, like this -

I propose 3 candidates -

  1. SteadyReturns
  2. Alpha1st
  3. kMini

Additionally, any strategy flagged in this manner should be removed from the rankings. Something ranked in the top 10 could be enticing to an unwitting subscriber who is a new user of the platform.

1 Like

Boy you are a legend. You discovered the problems back in April. It’s June now these strategies did explode.

Common factor here - high win ratio. So easy to exclude, yet these are the favourites of newbie subscribers. Some of them even occupied top c2 rating. Additionally clicking the small leverage button will show the hidden risk behind. Compare average leverage with max leverage on system page - if max is significantly higher than average, thats another red flag.

Past blow-ups by the same strategy manager are another red flag:

collective2.com/performance/138725816

collective2.com/performance/138598030

collective2.com/performance/138924621

collective2.com/performance/139520529

collective2.com/performance/140279524

Thank you! These are indeed scary… seems like some kind of scheme designed specifically for C2.

Your comment indirectly implies that systems with small leverage cannot blow up.
Nothing could be further from the truth, even systems with no leverage at all can go bankrupt.

It is an indicator of possible higher risk of blowup. But Eva doesn’t state anything else except that. One shouldn’t assume lack of that indicator means lack of risk of blowup.

An analogy would be if you believe high short interest indicates rise in stock price coming up, then lack of high short interest doesn’t mean stock will go down, it just means that indicator and what it implies isn’t there.

2 Likes

It is true QuantTiger, and I agree, mathematically the risk of ruin increases rapidly when the trader puts a higher percentage of his trading capital at risk.

1 Like