Hey guys,I'm coming towards the end of my PhD and was after some advice on what my next move should be.Basically I'm considering two options:1. Join an algo trading/HFT prop shop2. Go it alone and do my own algo trading and build from there (I'd basically be starting in my parents' garage)If I can develop something good should I do before someone else does and thereby negate the effectiveness of my strategy? Or does waiting while I gain experience from working in a prop shop outweigh this by giving me insight how I should run things, give me contacts in the industry and build a better monetary base from which to start? Or, even more pragmatically, is the fairy tale of starting in your parents' basement a thing of the past nowadays - even with platforms like MetaTrader 4 and the like?Any advice from experienced people would be appreciated.

My advise is - decide yourself.

how much capital do you have ?

Hi and welcome.My experience is short, less than 2 years. However I think I can help a bit. I started from scratch, didn't have any capital, built upon a model from applied physics, did extensive backtests and built around 15 strategies out of it. I selected the most promising one and after 10 months of making sure what the model is about, did a roadshow in a small EU country. Some people liked my results, hired me and put 1M $. I trade equities in NYSE, performance is pretty good (I outperform SP index), my payment is not since they still consider me junior but also due to my country's poor GDP that affects all (almost all..) wages.What I want to say is that even after studying, coding, backtesting and executing live, you need input from more experienced people. Perhaps instead of 15 months I would have spent 10 before going live. On the contrary since I've learnt from my mistakes, I will never repeat them again. It's a risk to put significant % of your capital, many things can happen to.you and lose money in the beginning.

QuoteOriginally posted by: kovaHey guys,I'm coming towards the end of my PhD and was after some advice on what my next move should be.Basically I'm considering two options:1. Join an algo trading/HFT prop shop2. Go it alone and do my own algo trading and build from there (I'd basically be starting in my parents' garage)If I can develop something good should I do before someone else does and thereby negate the effectiveness of my strategy? Or does waiting while I gain experience from working in a prop shop outweigh this by giving me insight how I should run things, give me contacts in the industry and build a better monetary base from which to start? Or, even more pragmatically, is the fairy tale of starting in your parents' basement a thing of the past nowadays - even with platforms like MetaTrader 4 and the like?Any advice from experienced people would be appreciated.Algo trading and HFT are different things, although HFT utilises algo-trading. You also seem to allude to systematic trading (which is maybe what you are calling algo trading) which is different again.HFT: strategies that exploit latency based features of market data, where those latencies are of the order of telecommunications/processing lag, or of the order of information aggregation into the SIP.Algo-Trading: Aims to optimally establish or exit positions. i.e. not 'trading-strategies' in the sense of constructing investment theses, but rather trading-tactics, in the sense that the strategy might be to implement position x; but how do we establish that position? This is of importance when position x involves a large quantity of money (market impact) or information leakage (about the intended strategy) or when the fund/traders simply want to minimise trading costs. Algo-trading should work against (toxic) HFT (in theory) although we have all seen recently how brokers can take both sides of the game to maximise house profit (surprise, surprise).Systematic-Trading: Any trading streategy that can be encapsulated with an algorithm... Can be low frequency or high frequency and is not even necessarily computerised (although it almost always is). Note that systematic trading can be statistical, macro, fundamental, technical etc... HFT and algotrading are by their nature almost always statistical or based on simply heuristics (that seek to exploit loose statistical tendencies of market microstructure).You dont have to go the whole mile in one go. To approach anything like a professional setup you need to spend serious money on data and support. Then on top of that you have to make money from your trading... very difficult when you have no experience.Why not just trade a private account? You will almost certainly loose your money... but you will learn what works for you. Dont assume that it has to be algo/HFT just because that is what everyone else is doing. If you have your own small account you want to be looking for anomolies that these guys cant exploit, not jumping into their trades, whereupon you will probably be crushed. It is also my personal opinion that HFT could be regulated away.

Thanks for the quick responses guys. I guess I could hedge my bets; trade a smallish amount of money on an account using something like MT4 and work for a prop shop to gain experience and, probably more importantly make experienced contacts I could possibly call upon if I do start my own prop shop. It sounds like from people's experiences that to be successful I'll need a bigger bankroll than what I have and plenty of good people. Thanks for sharing - it's given me a clearer picture of how to go about seeing myself up

QuoteI'm coming towards the end of my PhD and was after some advice on what my next move should be.What's the PhD on? 90% of all start up companies fail in the first 2 years. Quote It's a risk to put significant % of your capital, many things can happen to.you and lose money in the beginning. Do you relish having to do everything yourself (from marketing -> .... -> filling tax forms -> invoice reminders -> no social life).

ppauper's question about capital goes deeper than you might think...For pure latency arbitrage and the bleeding edge of HFT it is often better to model the risk not as "market risk" ie the chance that your bets will goes tits up, but "project risk", ie you are spending real money on hardware, programmers, quants, strats etc, and this investment may not pay off.As we have seen there are a large number of "market opportunities" that can only be exploited by having a cosy relationship with the exchanges, ie be a big customer.So there is a time structure to the game, with various conditions.Firstly you may build something good enough to grow into a business, the odds are against you, but the payoff is potentially good.The more likely path is that it doesn't fly. There are many modes from this, such as simply your model not working in practice, the market changing, it requiring more capital than you can access or the size of the exloit not being all that big.So if you do this, you must prepare to fail well....As you build this system, think of it as being just like your PhD or any other substantial project that you've done, in terms of not just being good, but looking good to someone who knows this stuff.If you fail, the next best thing is to turn up at a bank or HF and use it to show that you know this stuff and that you've learned harsh lessons that they benefit from but haven't shared your pain.So for instance, a bit more attention to risk and limit management, objective analysis of when to walk and when to run out of a position, recording in a journal every thing you've learned about the msarket you've targetted.The difference between a veteran and a victim is what you learn from your pain.It will be hard, that of itself can be spun when you go for a wage slave job, employers like worth ethic, being a self starter and a fuzzy thing I'm going to call "entrepreneurship".

Does MT4 offer a proper API to do proper algotrading? Interactive Brokers have a quite advanced trading platform.Another way to go is to start trading on your own, pimp your Sharpe ratio (which doesn't always mean maximising your profits...). If you don't get bored after a couple of months and your Sharpe ratio is consistently above 2-3, you can put your performance indicators on your CV and send it to some good HFs. I personally find equity market, researching and managing my equity portfolio much more exciting - you could try this too.a

or just invest in market trackers...a

This chart is not convincing, or is intentionally misleading if worded another way... kcg's loss is due to the merger (debts, managemental chaos, etc) in Jul 2013, its return gets much better in 2014.You should include 2014's data into your chart.QuoteOriginally posted by: katastrofaor just invest in market trackers...a

QuoteOriginally posted by: septfleurThis chart is not convincing, or is intentionally misleading if worded another way... kcg's loss is due to the merger (debts, managemental chaos, etc) in Jul 2013, its return gets much better in 2014.You should include 2014's data into your chart.QuoteOriginally posted by: katastrofaor just invest in market trackers...aAre you saying that http://www.bloomberg.com/news/2012-08-1 ... tware.html was accounted for in Q4 2013?a