Should I keep trading this system?

[Insurinator]

A "machine" will never trade as efficiently as a human for an extended period of time. Maybe in the year 2080 with all the supercomputers they will have around...

All the "successful" automated trading system's account statements I've ever seen show massive draw downs beyond any normal reasoning.

I would stop using it. Never should of touched it in the first place.

An automated system might be able to make returns for a couples months, possibly beyond a year's time... eventually the thing will turn around and rape you without mercy.

Just a warning.

 

[cgeorgan]

Quote from Insurinator:

A "machine" will never trade as efficiently as a human for an extended period of time. Maybe in the year 2080 with all the supercomputers they will have around...

All the "successful" automated trading system's account statements I've ever seen show massive draw downs beyond any normal reasoning.

I would stop using it. Never should of touched it in the first place.

An automated system might be able to make returns for a couples months, possibly beyond a year's time... eventually the thing will turn around and rape you without mercy.

Just a warning.

Interesting. I'm always drawn to words like "never", "always", "forever", etc.

Do share with us exactly why you made the statements you made. After all, my algorithms (there are more than one) didn't invent themselves. In fact, they were created to take advantage of specific human behaviors.

From experience (because how else can you program these things if you lack it), human traders engage in all sorts of interesting behavior:

1.)People love to look for trends and will hop on one when they find it.

2.)People love to bet on market swings and will jump on one when they find it.

3.)People can usually take a few good trades but, once one or two don't work out, their psyche is sufficiently degraded as to prevent their effective entrance into subsequent trades.

Regardless, nobody is going to show you their successful black boxes, which in my opinion should account for what you've "seen".

To the OP - this poster did say one useful thing: Systems don't last forever. Then again, human traders don't either.

 

[Insurinator]

Quote from cgeorgan:

Do share with us exactly why you made the statements you made. After all, my algorithms (there are more than one) didn't invent themselves. In fact, they were created to take advantage of specific human behaviors.

3.)People can usually take a few good trades but, once one or two don't work out, their psyche is sufficiently degraded as to prevent their effective entrance into subsequent trades.

Regardless, nobody is going to show you their successful black boxes, which in my opinion should account for what you've "seen".

To the OP - this poster did say one useful thing: Systems don't last forever. Then again, human traders don't either.

I made those statements because I can guarantee there are more successful traders to "black boxes". Just as your algorithms are created to take advantage of human behaviors there are human traders who will take advantage of other human behaviors too. It's a two way street, I just believe a trader who knows what they are doing is better than a machine at doing so. A good trader shouldn't be letting loses effect his traded either. That's not to say he isn't psychologically effected, but it should not effect his trade strategy significantly.

#3 is the exact problem to begin with when it comes "black boxes"... they simply don't give a damn. Hence the successful black boxes that I have seen turn profits for 1 - 2+ years and then all show 50% - 85% drawdowns. A human would never continue with a 85% drawdown because they would realize their strategy is severely flawed, rather than a machine with a flawed system ignoring the %85 drawdown and continuing to trade till bust.

I've never seen a legitimately successful blackbox or been offered one. I'de never touch the things. The successful one's that I've seen proof for though show rediculous drawdowns.

Sure the name of the game is to turn profits regardless which the successful black boxes that I have seen proof of do, but at a 50% - 85% drawdown that is not successful to me. It's a raping waiting to happen.

A human trader doing his due diligence and placing naked shorts is more likely to survive trading for a longer period of time with more profits. That is just my opinion though.

I do believe black boxes can be successful to a degree. I just do not believe they will outperform and have the same endurance as a proficient human trader.

 

[cgeorgan]

I'm going to respond to what you've written here, in the hopes that you walk away with what I say a with a little bit more knowledge than you had before.

Quote from Insurinator:

I made those statements because I can guarantee there are more successful traders to "black boxes".

This is a spurious argument. We are not discussing how many algorithms exist, or how many traders exist. We haven't even discussed "success", as it relates to "how successful", "length of success", etc. I find it tough to guarantee anything in this business - but if there's one thing I can guarantee, it's that you can't prove what you just said and - if you could - use it to create a cohesive argument either for or against human traders vs. algorithms.

Just as your algorithms are created to take advantage of human behaviors there are human traders who will take advantage of other human behaviors too. It's a two way street, I just believe a trader who knows what they are doing is better than a machine at doing so. A good trader shouldn't be letting loses effect his traded either. That's not to say he isn't psychologically effected, but it should not effect his trade strategy significantly.

This does not have to be a question of who is "better". Even if it were, we would have to compare apples to apples. Is a human better than an algorithm when trading 50 different instruments on a millisecond timeframe simultaneously? Furthermore, when talking about "better", to what are we referring - risk adjusted returns? Drawdowns? Profit factors? Win ratio? Win %?

#3 is the exact problem to begin with when it comes "black boxes"... they simply don't give a damn. Hence the successful black boxes that I have seen turn profits for 1 - 2+ years and then all show 50% - 85% drawdowns. A human would never continue with a 85% drawdown because they would realize their strategy is severely flawed, rather than a machine with a flawed system ignoring the %85 drawdown and continuing to trade till bust.

The question here is, again, not black boxes vs. humans. Black boxes are merely human tools - so in the case of the black box that has an 85% drawdown (again another random number), that is merely a human who ignores an 85% drawdown. This does not jive with your statement that "no human" would ever continue with an 85% drawdown.

I've never seen a legitimately successful blackbox or been offered one. I'de never touch the things. The successful one's that I've seen proof for though show rediculous drawdowns.

Which is it? If you've never seen a successful black box, how did you find a successful one with a ridiculous drawdown? Not only that, but how can you call it "successful" if it has a "ridiculous" drawdown?

You're right - no person in his right mind would ever offer you a successful algorithm, or even a piece of a successful algorithm. Some of the stuff I've created has a lot of moving parts, and I wouldn't even give you one of those parts - let alone the whole thing.

Success costs money. If somebody offered you a truly successful black box, you'd have to pay the present value of a certain # of years of profit. It's the same in the business world - you can't buy a successful business without paying both a discounted cash flow amount, as well as a premium to the owner. If the black box you're being offered is $19.99, you can rest assured that $19.99 is the present value of the cumulative profit its owner believes it will make at least in the next 10 years. Time to move on.

Sure the name of the game is to turn profits regardless which the successful black boxes that I have seen proof of do, but at a 50% - 85% drawdown that is not successful to me. It's a raping waiting to happen.

Again, which is it? If all black boxes you've seen show 50-85% drawdown, and you state that amount of draw down does not qualify something to be a "success", then how can you claim to have seen a "successful" black box?

It's like saying "I've never seen a man fly, but the ones I have seen fly have all fallen straight to the ground."

A human trader doing his due diligence and placing naked shorts is more likely to survive trading for a longer period of time with more profits. That is just my opinion though.

Thankfully, you've qualified this as your opinion, so I'll leave it as such. I have personally tried to apply the rulesets I've used in discretionary trading to program my algorithms. If my algorithms fail, it's most likely because my due diligence was wrong and my assumptions were incorrect. This has nothing to do with my algorithms and everything to do with me as a trader and programmer.

I do believe black boxes can be successful to a degree. I just do not believe they will outperform and have the same endurance as a proficient human trader.

Since you've never seen a successful black box, what you're really doing is just guessing they might work. You're probably guessing some might work because - gasp - people actually have working algorithms. For example, we have working algorithms right here where I work, and we're continuously revising the old ones and thinking up new ones.

Anyways, this was never an argument about who or what is "better" (humans vs. algorithms). I would never attempt to make such a comparison, as the market is far too large, diverse and secretive to ever make a successful comparison.

I hope this helps.

 

[Insurinator]

Quote from cgeorgan:

This is a spurious argument. We are not discussing how many algorithms exist, or how many traders exist. We haven't even discussed "success", as it relates to "how successful", "length of success", etc. I find it tough to guarantee anything in this business - but if there's one thing I can guarantee, it's that you can't prove what you just said and - if you could - use it to create a cohesive argument either for or against human traders vs. algorithms.

There is no way I can prove it. I'm just saying I would put my life on it that there are more successful human traders (eg. survive in general and make consistant profits for 10 years+) than there are successful black boxes.

This does not have to be a question of who is "better". Even if it were, we would have to compare apples to apples. Is a human better than an algorithm when trading 50 different instruments on a millisecond timeframe simultaneously? Furthermore, when talking about "better", to what are we referring - risk adjusted returns? Drawdowns? Profit factors? Win ratio? Win %?

I'm talking about pure survival which eventually should amount to profits.

The question here is, again, not black boxes vs. humans. Black boxes are merely human tools - so in the case of the black box that has an 85% drawdown (again another random number), that is merely a human who ignores an 85% drawdown. This does not jive with your statement that "no human" would ever continue with an 85% drawdown.

That's the thing.. most people who use blackboxes figure it's the machine trading their streamlined strategy with 100% discipline that they themselves would not be capable of doing. They decide not to closely monitor the machine because it trades by itself, and in return when things go south, they go very south.

Which is it? If you've never seen a successful black box, how did you find a successful one with a ridiculous drawdown? Not only that, but how can you call it "successful" if it has a "ridiculous" drawdown?

I have seen successful ones because they have shown consistant profits with a high win/loss ratio and the wins are on par with the loss's in terms of $ per win, $ per loss. Some of the wins are much larger than the losses.

The thing is they all show multiple periods of time where the account was 50% down and more. It's a disaster waiting to happen. 500 $10 wins to 5 $1 losses for 2 consecutive years doesn't mean anything if your account had to drawdown 50% - 85% 5 - 10 times to do so.

Surviving is more important in relation to turning profits than a win/loss ratio with higher wins than losses, but risking the entire account to do so. You won't be trading anything when the account blows.

You're right - no person in his right mind would ever offer you a successful algorithm, or even a piece of a successful algorithm. Some of the stuff I've created has a lot of moving parts, and I wouldn't even give you one of those parts - let alone the whole thing.

I wouldn't want it for free. I don't trust the things worth a damn. Backtesting is near usless to me which is how most black boxes prove their worth. The funny thing is 99% of black boxes do not backtest tick for tick which is the only real way to backtest, if your going to do it at all. These 99% backtest using 1 minute to 1 hour bars which completely skews the backtest into a successful strategy. The thing hits the market and gets destroyed quick... seen it a few times.

Success costs money. If somebody offered you a truly successful black box, you'd have to pay the present value of a certain # of years of profit. It's the same in the business world - you can't buy a successful business without paying both a discounted cash flow amount, as well as a premium to the owner. If the black box you're being offered is $19.99, you can rest assured that $19.99 is the present value of the cumulative profit its owner believes it will make at least in the next 10 years. Time to move on.

I would never make a down payment on one... way too much risk for the person purchasing it with 0 risk to the seller. If the seller used the system with his own money successfully like a real business owner would of used their own money to start the business or buy it, then yes in that case it would be a legitimate sale worth a down payment.

If none of their own money was used to profit with the system, I would tie equity into an account so I could use the system and pay the designer that way. I'de also make them put money worth 50% of my equity into an escrow account for 5 years incase any issues arise and my account goes down 50%. This would be of course if they tried to sell me the system flat out without ever using their own money on it first. This should be espcially easy for them to agree to because if offered a percent of the profits for the life that I apply the system it would potentially be an exponential amount of profits for them.

If they were confident in their system they should have no issue either way.

Again, which is it? If all black boxes you've seen show 50-85% drawdown, and you state that amount of draw down does not qualify something to be a "success", then how can you claim to have seen a "successful" black box?

It's like saying "I've never seen a man fly, but the ones I have seen fly have all fallen straight to the ground."

That's the thing... I have seen successful ones that turn profits and have not blown. But at consistant drawdowns of 50%+ it is bound to happen. Would you risk running the thing and having it bust before you make 10% ROI on it? You're just as well off throwing your money down on red at a roulette wheel.

Since you've never seen a successful black box, what you're really doing is just guessing they might work. You're probably guessing some might work because - gasp - people actually have working algorithms. For example, we have working algorithms right here where I work, and we're continuously revising the old ones and thinking up new ones.

I have seen successful ones and I do not doubt that you have working ones. For reasons I've made though, that is why you are revising them... because they likely will not last forever. Sure your the dynamic human designing it but the static machine is the one making the trades with money that is lost or won. To a degree, sometimes revision in trading can never come soon enough, and when it doesn't happen soon enough that's when you will feel it.

Anyways, this was never an argument about who or what is "better" (humans vs. algorithms). I would never attempt to make such a comparison, as the market is far too large, diverse and secretive to ever make a successful comparison.

Like I said before. I'de put my life on it that there are more successful human traders than there are black boxes that make the same amount. For all the human traders that would have an equally successful blackbox counterpart, there are and excess of human traders that are smashing the blackboxes.

 

[Insurinator]

Quote from cgeorgan:

This is a spurious argument. We are not discussing how many algorithms exist, or how many traders exist. We haven't even discussed "success", as it relates to "how successful", "length of success", etc. I find it tough to guarantee anything in this business - but if there's one thing I can guarantee, it's that you can't prove what you just said and - if you could - use it to create a cohesive argument either for or against human traders vs. algorithms.

There is no way I can prove it. I'm just saying I would put my life on it that there are more successful human traders (eg. survive in general and make consistant profits for 10 years+) than there are successful black boxes.

This does not have to be a question of who is "better". Even if it were, we would have to compare apples to apples. Is a human better than an algorithm when trading 50 different instruments on a millisecond timeframe simultaneously? Furthermore, when talking about "better", to what are we referring - risk adjusted returns? Drawdowns? Profit factors? Win ratio? Win %?

I'm talking about pure survival which eventually should amount to profits.

The question here is, again, not black boxes vs. humans. Black boxes are merely human tools - so in the case of the black box that has an 85% drawdown (again another random number), that is merely a human who ignores an 85% drawdown. This does not jive with your statement that "no human" would ever continue with an 85% drawdown.

That's the thing.. most people who use blackboxes figure it's the machine trading their streamlined strategy with 100% discipline that they themselves would not be capable of doing. They decide not to closely monitor the machine because it trades by itself, and in return when things go south, they go very south.

Which is it? If you've never seen a successful black box, how did you find a successful one with a ridiculous drawdown? Not only that, but how can you call it "successful" if it has a "ridiculous" drawdown?

I have seen successful ones because they have shown consistant profits with a high win/loss ratio and the wins are on par with the loss's in terms of $ per win, $ per loss. Some of the wins are much larger than the losses.

The thing is they all show multiple periods of time where the account was 50% down and more. It's a disaster waiting to happen. 500 $10 wins to 5 $1 losses for 2 consecutive years doesn't mean anything if your account had to drawdown 50% - 85% 5 - 10 times to do so.

Surviving is more important in relation to turning profits than a win/loss ratio with higher wins than losses, but risking the entire account to do so. You won't be trading anything when the account blows.

You're right - no person in his right mind would ever offer you a successful algorithm, or even a piece of a successful algorithm. Some of the stuff I've created has a lot of moving parts, and I wouldn't even give you one of those parts - let alone the whole thing.

I wouldn't want it for free. I don't trust the things worth a damn. Backtesting is near usless to me which is how most black boxes prove their worth. The funny thing is 99% of black boxes do not backtest tick for tick which is the only real way to backtest, if your going to do it at all. These 99% backtest using 1 minute to 1 hour bars which completely skews the backtest into a successful strategy. The thing hits the market and gets destroyed quick... seen it a few times.

Success costs money. If somebody offered you a truly successful black box, you'd have to pay the present value of a certain # of years of profit. It's the same in the business world - you can't buy a successful business without paying both a discounted cash flow amount, as well as a premium to the owner. If the black box you're being offered is $19.99, you can rest assured that $19.99 is the present value of the cumulative profit its owner believes it will make at least in the next 10 years. Time to move on.

I would never make a down payment on one... way too much risk for the person purchasing it with 0 risk to the seller. If the seller used the system with his own money successfully like a real business owner would of used their own money to start the business or buy it, then yes in that case it would be a legitimate sale worth a down payment.

If none of their own money was used to profit with the system, I would tie equity into an account so I could use the system and pay the designer that way. I'de also make them put money worth 50% of my equity into an escrow account for 5 years incase any issues arise and my account goes down 50%. This would be of course if they tried to sell me the system flat out without ever using their own money on it first. This should be espcially easy for them to agree to because if offered a percent of the profits for the life that I apply the system it would potentially be an exponential amount of profits for them.

If they were confident in their system they should have no issue either way.

Again, which is it? If all black boxes you've seen show 50-85% drawdown, and you state that amount of draw down does not qualify something to be a "success", then how can you claim to have seen a "successful" black box?

It's like saying "I've never seen a man fly, but the ones I have seen fly have all fallen straight to the ground."

That's the thing... I have seen successful ones that turn profits and have not blown. But at consistent drawdowns of 50%+ it is bound to happen. Would you risk running the thing and having it bust before you make 10% ROI on it? You're just as well off throwing your money down on red at a roulette wheel hoping you don't bust too early.

Since you've never seen a successful black box, what you're really doing is just guessing they might work. You're probably guessing some might work because - gasp - people actually have working algorithms. For example, we have working algorithms right here where I work, and we're continuously revising the old ones and thinking up new ones.

I have seen successful ones and I do not doubt that you have working ones. For reasons I've made though, that is why you are revising them... because they likely will not last forever. Sure your the dynamic human designing it but the static machine is the one making the trades with money that is lost or won. To a degree, sometimes revision in trading can never come soon enough, and when it doesn't happen soon enough that's when you will feel it.

Anyways, this was never an argument about who or what is "better" (humans vs. algorithms). I would never attempt to make such a comparison, as the market is far too large, diverse and secretive to ever make a successful comparison.

Like I said before. I'de put my life on it that there are more successful human traders than there are black boxes that make the same amount. For all the human traders that would have an equally successful blackbox counterpart, there are and excess of human traders that are far surpassing the blackboxes.