Maximun Drawdowns

Browsing systems… Noticed a system that is up 154% since inception, with a 45% annual return, started with $100,000 and never dipped below $90,00 in system equity and now has $254,000, and yet according to the listed stats it has a 99 percent drawdown.

How can that be?

Check out the Peak to Valley drawdown. Assume someone would have subscribed at the “Peak”… they would have lost 99% when they reached the “Valley.”

Thanks for the explanation. I checked the peaks to valleys in the system referenced. The C2 numbers are obviously inaccurate because of a couple errant ticks that are calculated into the draw down numbers (and presumably in the developer’s rating also).

In one minute on 1/24/14 the system went from a net profit of $246K to $370K and back to $239K. That’s a 99% draw down right in that one minute, but given that it was solely in an ETF, that is obviously an errant tick. It happened again on 4/29/14 when it went from a net profit of $232K down to $117K and back up $236K in, it appears again, one minute. Outside of those two ticks I can not find a draw down greater than 29 percent who knows if even that is accurate?.

Does make me wonder how often errant ticks work in the favor of percentage gains in this or that system.

Hypothetical results, indeed.

Drawdown should perhaps be calculated based on close to close rather than on tick (or even intraday H/L).

It would be helpful if you told me the actual system ID so I can remove the bad tick.

Actually, it would be more helpful for everyone if you addressed this issue systematically. I suspect there are bad ticks all over this site.

A quick search of the grid shows there are 46 systems with max draw downs of more than 80%. Twenty-five of those have draw downs of 100%. Eleven of the 46, despite the 80% draw downs, have annual gains for !00% or more.

And it’s not only in the draw downs, there are 51 systems here making 200% or more annually. That’s absolutely great if they are actually doing those numbers but those kinds of returns are at least questionable if not utterly bogus.

Maybe it’s just the grid numbers that are screwed up but it seems there may be a pretty big data problem here and not a lot of due diligence going on about it.

And I have to ask – are these numbers part of what’s used to determine developers’ ratings? If so, the rating system either needs to be revamped or eliminated. If it’s to be a revamping, Mr. McGrath’s suggestion that tick not be used for anything here is probably a good place to start.

close to close will not show you what drawdown is supposed to show you. its exactly what the name implies. C2 has it correct and that’s how drawdown works.

I think Michael is talking about his own system. Not sure why you won’t say that in the first place. If it’s a bad tick, have Matthew fix it.