Looking for the Worst Automated Systems

Random.Capital · Feb 13, 2009

Quote from Craig66:

Yeah, that always works awesome.

I think folks forget sometimes that taking the "anti-trade" isn't just about taking a position in the opposite direction, it's about having the inverse stop as well. So if the losing system is taking a long with a relatively tight stop, the anti-trade is to go short with a monstrous stop.

It's not just an inverse position - it's an inverse style of trading.

Euler · Feb 13, 2009

Quote from eagle:

The most misunderstand and the most important is the word "consistent". Designing a consistent winning system or a consistent losing system is the same thing.

I think this is true only in markets with:

1. 0 transaction costs
2. infinite liquidity

which doesn't exist, although some markets may come close to it (e.g., ES).

uexkuell · Feb 13, 2009

Ok, here is some incredibly bad system (and I give it away for free)

:

Just do buy@market with a very high frequency and then sell@market as fast as your hardware/software/broker allows you.

Choose a future with a consistently huge spread for this.
Have your account with a broker that gives you a margin of 500$ per car.

For a quick computation of the return let's assume ES (not optimal since it has a tight spread, anyway..).
In one round-turn you will mostly have a loss of 25$+commissions.
If you manage to have only one round-turn every second you can loose 90k (commissions not included) in one hour with only 500$ cash.

Telling about "my system" only to show the danger of inversing a loosing system. I think many loosing systems do have some aspect of the above mentioned in it.

shortie · Feb 13, 2009

Quote from uexkuell:

Ok, here is some incredibly bad system (and I give it away for free) :

Just do buy@market with a very high frequency and then sell@market as fast as your hardware/software/broker allows you.

Choose a future with a consistently huge spread for this.
Have your account with a broker that gives you a margin of 500$ per car.

For a quick computation of the return let's assume ES (not optimal since it has a tight spread, anyway..).
In one round-turn you will mostly have a loss of 25$+commissions.
If you manage to have only one round-turn every second you can loose 90k (commissions not included) in one hour with only 500$ cash.

Telling about "my system" only to show the danger of inversing a loosing system. I think many loosing systems do have some aspect of the above mentioned in it.

the broker uses the inverse of your system to consistently make money. so inverting a losing system is a valid approach. one just needs to know to invert it correctly.

the market maker also uses a reversed version of your system and also makes money.

4XQs · Feb 13, 2009

Quote from Random.Capital:

I think folks forget sometimes that taking the "anti-trade" isn't just about taking a position in the opposite direction, it's about having the inverse stop as well. So if the losing system is taking a long with a relatively tight stop, the anti-trade is to go short with a monstrous stop.

It's not just an inverse position - it's an inverse style of trading.

But: If the system has a Profit Target then it's bound to get out somewhere - that would be your stop. I think it would depend on how the system defines it's PT - if it's the kind of "let your profits run and wait for a signal in the opposite direction" then you can very well go bankrupt waiting for that to happen on that outlier trade.

EDIT: I really should mention that this thread absolutely cracks me up

eagle · Feb 13, 2009

I meant if you can find one no matter it is a winning or a losing system but make sure it is consistent then you find the Holy Grail.

Quote from Euler:

I think this is true only in markets with:

1. 0 transaction costs
2. infinite liquidity

which doesn't exist, although some markets may come close to it (e.g., ES).

Euler · Feb 13, 2009

Quote from shortie:

the broker uses the inverse of your system to consistently make money. so inverting a losing system is a valid approach. one just needs to know to invert it correctly.

the market maker also uses a reversed version of your system and also makes money.

That depends on how you define "inverse", because the broker or market maker's counterparty may be far more informed, whereas uexkuell's strategy is totally is uninformed; hence, brokers and/or market makers often do lose money, and even go bankrupt, whereas those who trade against (the fictional) uexkuell strategy would, with asymptotically certain probability, make money.

The so-called "opposite" of a losing strategy is often another losing strategy once asymmetry of information, rationality of participants, and transaction costs are figured in.

andrewbee · Feb 13, 2009

Yes, transaction costs and slippage (unless opening limits are used) are constants that drag down the performance of any system, even when you invert it. The lower the frequency of trading, the less they affect it.

BTW, I was totally serious about inverting this system (it does use fixed stops and limits, so these would naturally be inverted too). I would not do so without 6 months of trading history to go off of. So far I only have 2 weeks though, that cover what has been an atypically volatile period. ($%#% stimulus bill bollocks. The best way the gov't could stimulate the market is by SHUTTING THE HELL UP!).

intradaybill · Feb 13, 2009

So this is my conclusion from this very intersting thread:

Since a high percentage of traders lose, their combined startegy is a losing one. The inverse startegy of their combined startegy is the winning one.

Therefore, the winning startegy is based on doing the opposite of what the majority of traders is doing at any given time.

hmmm...

harooki2 · Feb 13, 2009

By George, this has turned into an interesting thread. I might benefit from the discussion as much as any losing system someone might send me.

There are many complications in turning a system upside down. Among them are spread, time frame, and trade frequency. If the spread is too large, it will eat you up regardless of which way you go. The shorter time frames usually make it impractical to turn a system upside down, because of the trade frequency. Therefore, you have two factors that collide.

The bottom line is that it ain't as easy as finding a losing system and mirroring it in the opposite direction. The other factors that you must consider come from experience, the best of all teachers.

That said, it is inevitable that taking the other side of the trade can be profitable. But eagle hit it on the head when he said that "if you can find one no matter it is a winning or a losing system but make sure it is consistent then you find the Holy Grail." A losing system that is losing inconsistently (some good days, some bad, some good weeks, some bad, etc.) is not a good candidate. If you fade that type of system, you'll wind up in the same boat that got you here.

It is inarguable that the other side of most amateur traders' trades is the profitable side. Otherwise, they wouldn't be losers. So, who can argue, then, that getting on that other side is the wrong approach?

Someone else had a very good argument for a slightly differenct approach. That was to find a trader who is influenced by fear and greed and fade him/her. I think that there is surely merit in that, but it is impractical. The next best thing is to fade the prevailing logic among amateur traders.

I've gone on too long. I'll leave it at that for now.