II blames style-drift for CTA woes

[Maverick1]

I am employed in the industry myself and one of the aspects of managing a portfolio involves keeping tabs on the competition. My friends are working all over the street - while nobody would give up their trade secrets (nor would I), I have a general idea about who is doing what.

I’ll tell you a scary secret - there is no secret sauce, it’s the little differences in the recipe and the cooking process that makes the difference.

I believe you Santa, re lil' things... the devil is always in the details...

That's precisely one of the key reasons why I expect the quant bandwagon at large to fail; the kind of mind it takes to decipher which details matter is not developed in the halls of a computer science department, no matter how prestigious

Think of it like this: there were plenty of engineers that could build a phone, but there was only one Steve Jobs... It's a bit like that in the markets

Hard to believe you re Rentec even if it's a general idea, but will take you at your word

 

[helpme_please]

I believe you Santa, re lil' things... the devil is always in the details...

That's precisely one of the key reasons why I expect the quant bandwagon at large to fail; the kind of mind it takes to decipher which details matter is not developed in the halls of a computer science department, no matter how prestigious

Think of it like this: there were plenty of engineers that could build a phone, but there was only one Steve Jobs... It's a bit like that in the markets

Hard to believe you re Rentec even if it's a general idea, but will take you at your word

Hi Maverick1,

Renaissance Technologies have a recruitment policy of hiring the best mathematicians/scientists who have a track record of doing good science/math. Note, good science/math and not good finance/trading. So, you're hiring a bunch of brilliant people with no domain knowledge in investing/trading, yet they beat the best domain experts at their own game through their quant skills. So, this is one rare occasion when I don't agree with you. I don't think that the quant bandwagon at large will fail. I will be happy if someone tells me why I'm wrong.

 

[helpme_please]

I am employed in the industry myself and one of the aspects of managing a portfolio involves keeping tabs on the competition. My friends are working all over the street - while nobody would give up their trade secrets (nor would I), I have a general idea about who is doing what.

I’ll tell you a scary secret - there is no secret sauce, it’s the little differences in the recipe and the cooking process that makes the difference.

Hi Secret Santa,

Thanks for sharing your scary secret. I believe the broad principles are out there in books and even here on this forum. I do not work in this industry. May I ask how how frequently do the insiders in this industry tweak the recipe and cooking process? If it is as infrequent as 2 times out of 10 years, wouldn't many people working in this industry be jobless? Why not simply leave most of the work to machines as machines make less mistake and are very cheap once they are properly programmed?

 

[sle]

May I ask how how frequently do the insiders in this industry tweak the recipe and cooking process? If it is as infrequent as 2 times out of 10 years, wouldn't many people working in this industry be jobless? Why not simply leave most of the work to machines as machines make less mistake and are very cheap once they are properly programmed?

It's a constant process - managing risk allocations, deciding to turn alphas/strategies off, researching new alphas etc. Stuff that was working 5 years ago does not work today and stuff that works today is probably going to be obsolete in a year.

 

[Van_der_Voort_4]

Hard to believe you re Rentec even if it's a general idea, but will take you at your word

If someone went to their office and told them what he thinks they are doing and why he came to this conclusions and was correct,the look on their faces would be priceless.I suspect only few people at the very top know the actual strategy itself,even if general work environment is to share and collaborate.And i am not implying they are doing anything illegal,just that they are very,very clever.

Not that i do know myself,far from it,i only have a theory like everyone else.

 

[Maverick1]

Hi Maverick1,

Renaissance Technologies have a recruitment policy of hiring the best mathematicians/scientists who have a track record of doing good science/math. Note, good science/math and not good finance/trading. So, you're hiring a bunch of brilliant people with no domain knowledge in investing/trading, yet they beat the best domain experts at their own game through their quant skills. So, this is one rare occasion when I don't agree with you. I don't think that the quant bandwagon at large will fail. I will be happy if someone tells me why I'm wrong.

I exclude Rentec from the rest of the pack. Simons and crew will find a way to avoid getting killed in the downturn that's coming. The others won't. Underperformance has already started, so get woke to reality

 

[ironchef]

It's a constant process - managing risk allocations, deciding to turn alphas/strategies off, researching new alphas etc. Stuff that was working 5 years ago does not work today and stuff that works today is probably going to be obsolete in a year.

If they have to constantly tweak their system to make them work, could they all be fooled by randomness?

 

[ironchef]

only thing that is sustainable is fundamentals with a technical overlay. Anything short of that will fail longer term, including the recent 'quant' driven/AI/machine-learning rabbit hole

I agree if you have a longer time horizon than day trades.

My struggle with technical overlay is what should I be watching out for? Except perhaps Price, normal TA like RSI, MACD, MA... did not work for me.

Thanks.

 

[sle]

If they have to constantly tweak their system to make them work, could they all be fooled by randomness?

For most people with multiple high quality strategies/signals, the primary tweaks are more along the lines of seeing where the alpha has declined (potentially to the point where it stopped working) and scaling these strategies down. Then there is a matter of adding new alphas - same process, you start small and gradually allocate more risk. So that doesn’t make really make it random.

However, once you are living in the CTA or factor investing world. The alphas there are relatively few, really long-dated and noisy and usually the ones with large capacity are over-allocated. Understanding what works and what died becomes super tricky and I think a lot of science devolves into an art form (not necessarily bad, but it is not really the promised quantitative process).

 

[Maverick1]

I agree if you have a longer time horizon than day trades.

My struggle with technical overlay is what should I be watching out for? Except perhaps Price, normal TA like RSI, MACD, MA... did not work for me.

Thanks.

Re time frame, the sweet spot is days to weeks...

About technicals, whatever your coding/testing uncovers is what I meant. Think of it as a subset of your work with data