QuoteOriginally posted by: GamalIn algo trading you use all trading policies known from the "macro" world. Some are more scientific, some less.Look at the way traders talk. They've got precise trading algorithms but have serious problem in describing them. To extract all that from their minds and code it smartly - it's a real challenge. But still it's not maths related.This isnt really true, most statarb operates at much higher frequencies than human traders and hence the strategies are different; humans are not trading at the millisecond level nor are they looking for arbitrage strategies over hundreds of instruments simultaneously. I'm sure there are people out there who are trying to formalize 'traders intuition' but this is not typically what the big HFT funds are doing. I guess it depends what you mean by 'algo trading' though.

QuoteOriginally posted by: EriatarkaQuoteOriginally posted by: GamalIn algo trading you use all trading policies known from the "macro" world. Some are more scientific, some less.Look at the way traders talk. They've got precise trading algorithms but have serious problem in describing them. To extract all that from their minds and code it smartly - it's a real challenge. But still it's not maths related.This isnt really true, most statarb operates at much higher frequencies than human traders and hence the strategies are different; humans are not trading at the millisecond level nor are they looking for arbitrage strategies over hundreds of instruments simultaneously. I'm sure there are people out there who are trying to formalize 'traders intuition' but this is not typically what the big HFT funds are doing. I guess it depends what you mean by 'algo trading' though.You are confusing statarb and HFT here. Stat arb can describe any quant trading, but usually refers to some sort of pair trading, not necessarily high frequency. HFT is anything with very short holding times, usually market making. Anyway, quant funds are interested in whatever makes money and not the most sophisticated way to make money. If some "trader's intuition" works and it is possible to code it and test on hundreds of instruments they will do it. If they consider a PhD in machine learning useful they will probably interview the guy. But there are no magic machine learning formulas for making money.

QuoteOriginally posted by: EriatarkaI'm not sure where you got the idea that algotrading is about trying to formalize technical analysis.From my personal experience and talks to friends. Statarb is an important piece of algo trading but only a piece.If you want some science behind algo trading, yes, there's science. Namely fractals: what is good in macroworld is also good in microworld.Algo trading consists of course mostly of analysis of data. For the purpose of making money sorcery is as good as science.

QuoteOriginally posted by: qqqqqYou are confusing statarb and HFT here. Stat arb can describe any quant trading, but usually refers to some sort of pair trading, not necessarily high frequency. HFT is anything with very short holding times, usually market making. In practice they tend to be the same thing, most places doing HFT are doing some kind of statarb, and statarb is usually done at high frequencies. Also its hard to make a real distinction between market making and HFT, since HFT/statarb/pairs-trading is often just a sophisticated type of scalping and functionally identical to market making.There are also people doing algotrading which isnt statarb/HFT, for example some hedge funds use statistical analysis to look for alpha in longer term positions, and banks can use algotrading for automated market making in asset classes which arent liquid enough to warrant HFT (for example credit derivatives) But in practice if I heard a hedge fund was doing algotrading, I would usually think hft/statarb as a first guess.

.

QuoteOriginally posted by: EriatarkaQuoteOriginally posted by: qqqqqYou are confusing statarb and HFT here. Stat arb can describe any quant trading, but usually refers to some sort of pair trading, not necessarily high frequency. HFT is anything with very short holding times, usually market making. In practice they tend to be the same thing, most places doing HFT are doing some kind of statarb, and statarb is usually done at high frequencies. Also its hard to make a real distinction between market making and HFT, since HFT/statarb/pairs-trading is often just a sophisticated type of scalping and functionally identical to market making.There are also people doing algotrading which isnt statarb/HFT, for example some hedge funds use statistical analysis to look for alpha in longer term positions, and banks can use algotrading for automated market making in asset classes which arent liquid enough to warrant HFT (for example credit derivatives) But in practice if I heard a hedge fund was doing algotrading, I would usually think hft/statarb as a first guess.Most stat arb hedge fund strategies are not really high frequency. You cannot manage hundreds of millions to billions of dollars using HFT, the capacity of these strategies is very low. Most HFT are prop shops.

QuoteOriginally posted by: qqqqqMost stat arb hedge fund strategies are not really high frequency. You cannot manage hundreds of millions to billions of dollars using HFT, the capacity of these strategies is very low. Most HFT are prop shops.LTCM were HFT, Knight Capital were HFT, Gloucester Research, Rentech, DE Shaw and Winton are reported to use a lot of HFT, thats some of the biggest quant funds (of course, slightly more is known about the strategies of the places that blew up)

QuoteOriginally posted by: EriatarkaQuoteOriginally posted by: qqqqqMost stat arb hedge fund strategies are not really high frequency. You cannot manage hundreds of millions to billions of dollars using HFT, the capacity of these strategies is very low. Most HFT are prop shops.LTCM were HFT, Knight Capital were HFT, Gloucester Research, Rentech, DE Shaw and Winton are reported to use a lot of HFT, thats some of the biggest quant funds (of course, slightly more is known about the strategies of the places that blew up)I know that Winton is a trend-following CTA. It is hardly HFT, though they might have some strategies of this type.LTCM was active when HFT in its modern meaning was not possible.Rentech Medallion might use HFT, but it started trading when it was not available.

Considering the roles I've been contacted about at Winton by headhunters over the last few years they've clearly been trying to get into HFT more and more.

Lots of information here about algo trading. That's very helpfull but does only partially answer the initial question: how a derivative quants (5y+ exp) can move to algo trading? One idea would be to find a part-time MSc involving participating in a algo trading competition and get a good track record.

Derivatives quant know good amount of Math, but very little programming and data manipulation.Algo / HFT people are not so good in Math, but are expert in programming and data manipulation.Just fill in the gaps and then see if you are prepared for the algo quant interviews with multithreading, and other technical stuff. They prefer strong computer science people than Finance / Math people.

QuoteOriginally posted by: traderjoe1976Derivatives quant know good amount of Math, but very little programming and data manipulation.Algo / HFT people are not so good in Math, but are expert in programming and data manipulation.Just fill in the gaps and then see if you are prepared for the algo quant interviews with multithreading, and other technical stuff. They prefer strong computer science people than Finance / Math people.This depends what sort of job you're going for. If you are implementing production code then yes, you will need to be a great programmer but in that case you are a developer not a quant. However if you are doing algorithm research then programming isnt as important and you dont need to (eg) be a C++ expert like you do in derivatives. In the several quant fund interviews Ive had for these sort of roles I've never been asked programming questions beyond the basics, they are instead mainly related to mathematics, statistics and machine learning.

@ZhuliAn, Is this question out of personal choice or are things getting really tight for derivative quants?

Both

Market for exotic derivatives quant is very shallow now. It's better with flow derivatives.