I’ve asked a version of this question in another forum and only got a few responses–and those responses weren’t straight forward.
How long, would you say, it takes for a system to go from having abnormally high gains, to crash-and-burning?
…usually the minute I subscribe lol - yours being the exception thus far. Seriously, you can note on C2 it varies from a few weeks to a few years - there is no ‘one fits all’ . A few years ago forex systems were all the rage, this year it is volatility systems…who knows where next. What I have noticed more recently is that leading systems appear to be getting more robust i.e. better back-testing/sampling/hedging etc?
1 Like
6 months on average …
Lol…I’m approaching the 6 mo. mark. If I can make it through June, I guess I’ll be golden. 
LOL…thanks.
That’s what I’d assume too.
Natural selection?
“On average” , so even after a year you’re not off the hook …
Cool. I thrive when I’m on the hook. Under pressure. Hated on. Doubted.
I’m like that villain monster that gets stronger and bigger when the good guys shoot there weapons at it.
Seriously. That sort of thing motivates me and gets my brain juices flowing more than money can. More than anything else could.
Same here … maybe you should thank me then .
Where are my manners. Thank you.
1 Like
In my opinion system performance peaks just about the time a developer touts his system on C2 as the best ever, can do no wrong, and will continue to excel no matter what the markets do. Confidence is great but over-confidence will lead to sloppiness, mistakes, looking past warning signs, over trading, over optimizing, etc. JMHO
1 Like
Many times, a developer is put on the spot by the questions they are asked. When they answer in a manner they believe to be truthful, they are labeled as being over-confident. What does “over-confident” mean anyway? Is that a matter of opinion?
Anywho, it seems you’re suggesting that I reached that threshold, and as a result, my performance should begin to fall off?
“How long Does it take High flying systems to crash ?”
That usually happens when the market start to move out of the ordinary behaviour with a substantial increase of the volatility (and panic) in the markets aka market crisis.
It is the time where curve fitted systems miserably start to fail and model based (for instance) with a martingale money management approach, start to add contracts (and loss) for their followers. I guess one could check what happened to many great systems prior august 2015 or Jan 2016. However that were only fast events the whole picture one could get analyzing protracted crisis event such as 2008 or 2001. The common mistake is usually wrong expectation by the trader/investor.
1 Like
One problem is that certain systems do well in certain environments that may last several months or even a year or two. After that, the system fails as the environment shifts. For example, gold has been in a months-long up trend since the beginning of the year in which a number of gold-trading strategies have done very well. Will these same systems perform as well in a sideways market or downtrend market? Since we have yet to enter such times, it’s hard to say. One way of challenging a system in this regard is to backtest it against such periods in the past without curve fitting. Otherwise, a system that performs mightily in the current environment will suddenly suffer massive drawdowns that would be difficult to reverse as the system is not designed to work in the new environment.
Another problem is that certain developers purposefully make reckless bets at the beginning of the strategy life cycle to obtain very high returns in a short period (often done across several new strategies, several of which fail immediately and are thus discarded/hidden away from view) to attract subscribers then make much smaller bets overall to maintain the high return, sharpe ratio and drawdown percentage. They then milk subscribers for several months before such subscribers either (a) get impatient with terrible returns or (b) get annihilated due to additional risky bets that the developer makes to maintain high returns.
Another problem is that certain systems have high success but catastrophic hidden tail risk that is not readily quantified and explained upfront. For example, martingale strategies wherein developers double-down on losing trades in the hopes of a reversal will often have high win percentages, but will undoubtedly suffer catastrophic losses at a future point in time. Subscribers are unaware of this catastrophic tail risk since there is no measure that captures it. Another such strategy would be selling calls or puts wherein the premium received may not offset catastrophic adverse moves. Again, such tail risk is difficult if not impossible to quantify.
8 Likes
Truthful is good and as you suggest over-confidence is subjective.
Having experienced several systems which have crashed and burned I feel psychology is as important as technical skills and algo development skills and hope, for obvious reasons, you consider that facet of trading.
Best wishes for continued success!
1 Like
Nice answer.
I find that ‘curve-fitted’ systems fail much sooner than that. But I haven’t seen any complete life-cycles of systems here on C2 yet. So maybe they last longer here.
Zip4x is the only system I know of on C2 that has several-year track record with stellar returns. Several posters have noted, though, that after auditing trade-by-trade drawdowns there are several instances in which the system appears to go completely bust before returning to profitability in the same day. Whether this would be an account-busting event in real life is difficult to determine, and it has only occurred a few times in the strategy’s life. Nowadays, the system also seems to have tempered its betting strategy in favor of a more stable approach (at the cost of returns) now that it has established a very high compounded annual return statistic to attract subscribers.
Thank’s for the wishes, I’m over-confident that we’ll both (continue to) be pleased
Yes, I have psychological awareness. I spoke on it earlier.
Correlation is not necessarily causation. You noticed confident traders fail, so your conclusion is confidence makes traders fail.
IMO, there are confident good traders; and confident bad traders.
1 Like
Well, it’s obvious you know what you’re talking about. I appreciate you giving CkNN Algo a chance.
It just has to happen once that would be enough …
2 Likes
