Can you design a horribly losing system?

[1a2b3cppp]

I've been trying all night.

A good starting place and good thought exercise is to start with a few of your favorite popular indicators and reverse the signals.

You'll find you still do about the same.

Take something like stochastics that is supposed to tell you when the market is "overbought" or "oversold" and you're supposed to buy or sell there. If you reverse them (for example, sell when the indicator says it's oversold), you'll find that you end up making money during big trends, and losing money otherwise, which is basically the opposite of what happens when you follow that indicator's normal signals.

Neither is profitable, though.

Try to design a consistently losing system.

 

[Stoopidly]

I'd just create a system that make a lot of trades each day. My system would lose very consistently due to commissions and slippage. I guess thats cheating though.

From my experience (daily+ bars, fewer trades), creating a consistently losing system is just as hard as doing the opposite. Maybe slightly easier if you account for commissions and slippage. Given enough time/data it ends up in relatively break even, slightly winning or losing. With huge swings.

 

[Satan's Helper]

Any indicator system is random. Buy, sell or anything makes no sense. You can't prepare for today's weather with last years forecast!

X Hamster

 

[IronFist]

Quote from Stoopidly:

I'd just create a system that make a lot of trades each day. My system would lose very consistently due to commissions and slippage. I guess thats cheating though.

From my experience (daily+ bars, fewer trades), creating a consistently losing system is just as hard as doing the opposite. Maybe slightly easier if you account for commissions and slippage. Given enough time/data it ends up in relatively break even, slightly winning or losing. With huge swings.

Over trading to lose money on commissions probably doesn't count.

 

[Wide Tailz]

Here's a bad system:

Buy a 40 bar breakout with a .1 ATR stop.

:eek:

 

[Nab]

Quote from 1a2b3cppp:

I've been trying all night.

Try to design a consistently losing system.

Silly question. This is identical to finding a consistently winning system (as long as the loses do not mainly come from slippage/commissions).

 

[1a2b3cppp]

Quote from Nab:

Silly question. This is identical to finding a consistently winning system (as long as the loses do not mainly come from slippage/commissions).

Yeah, but it's weird when you actually try to make a losing system.

 

[1a2b3cppp]

Quote from Wide Tailz:

Here's a bad system:

Buy a 40 bar breakout with a .1 ATR stop.

:eek:

Can you post a chart showing that?

 

[Wide Tailz]

In order to flip these signalz around to get a profitable system, there are a few additional considerations, but there is at least one guy in the ES journal doing just this, based on the equity curve he posted.

It's just crude mean reversion: fade all breakouts with a tiny target, because suckerz always buy them and get slammed.

 

[Nab]

Quote from 1a2b3cppp:

Yeah, but it's weird when you actually try to make a losing system.

Depends on the time-frame. Clearly a consistent short strategy on a longer time-frame is more difficult to find than a long strategy. If the time-frame is small enough, its becomes equally difficult ...