I think it would be hard. Quite hard. I think only about 20% of my automated stuff would work with an IB fee structure and no exchange memberships. There might be more opportunities in equities or esoteric products I'm not familiar with or don't trade. But I focus on strategies that trade a lot. There could very well be some medium term swing stuff that could work. I've seen some of that. I'd focus on one product, generate some volume, and then move off IB and see what discounts you can get via exchange memberships or volume rebates. I've only created 2-3 threads here. In one of them I talk about how I made the transition.
Automated Trading - FAQs
- Thread starter Alpha Trader
- Start date
I think it would be hard. Quite hard. I think only about 20% of my automated stuff would work with an IB fee structure and no exchange memberships. There might be more opportunities in equities or esoteric products I'm not familiar with or don't trade. But I focus on strategies that trade a lot. There could very well be some medium term swing stuff that could work. I've seen some of that. I'd focus on one product, generate some volume, and then move off IB and see what discounts you can get via exchange memberships or volume rebates. I've only created 2-3 threads here. In one of them I talk about how I made the transition.
Garachen,have you ever tried to discard the data series and use certain events only?As an example of dumping data series,''trade the news'' would suffice.In a common sense,it`s still data series,but as the same time it is not,if you know what i mean.
The following link goes to answering one of the questions from the original post: What is/are the best language to learn for setting up trading systems?
The response is explained well, in a simplistic fashion and also divelges into the reasoning (benefits/specificities) for each language.
There are further links that expand into various directions and are helpful in answering basic questions.
http://www.quantstart.com/articles/...Should-You-Learn-To-Get-A-Quant-Developer-Job
The response is explained well, in a simplistic fashion and also divelges into the reasoning (benefits/specificities) for each language.
There are further links that expand into various directions and are helpful in answering basic questions.
http://www.quantstart.com/articles/...Should-You-Learn-To-Get-A-Quant-Developer-Job
Since you mentioned that only 20% of your algo strategies would even be profitable on a retail platform, it seems like a catch-22 for retail traders. For traders who are interested in ATS development, even smart ones who are dedicated and have lots of time to devote but doesn't have a brother that works at Getco, it seems like it would be much more optimal to develop strategies on an institutional platform considering many strategies wouldn't work on a retail platform. However, he/she would already need a level of success already, in order to invest $, make volume commitments, etc. to get access to better fees/execution on an institutional platform.
Do you have any recommendations for retail traders who are interested in ATS development as far as educational resources, open-source projects, common misconceptions, things to avoid, etc.. specifically those that are on a retail platform? I'd love to dedicate more time to investigating ATS, but even though I trade equities, I suspect most ATS edge would be microstructure-dependent and as a retail trader I don't want to head down a rabbit hole knowing if I actually had the background/capital/fee structure I'd be focusing on completely different things.
Last edited:
Let`s solve the riddle!
Look how many 7s in the table and only one 12.If you throw dices,the chance to get 7 is VERY high,whereas hiting the 12 is very low.Now,applying it to the market environment,would it be similar to set 7 ticks for profit and 12 tiks for stop targets?What studies are there to get the similar distribution for the market hours?Maybe Garachen and alike will chime in and give some tips.
Those are all good points about things that I see that I've been concerned about lately. Many of the large trading firms now got started in the early 2000s. Back then, a little bit of automation went a long way and it was pretty easy to bootstrap your way into success. I started in 2006 (a bit late to the game). I was lucky that there were a few exchanges/products with no automation and so I was the first there. That, in large part provided the capital to bootstrap myself into more competitive products. I still see a few US based products that are juicy enough for a new ATS to mess with but no so juicy that any of the big shops care. But it's not like the early 2000s where opportunities were everywhere.
To be clear. I'm talking from the standpoint of US futures. There's always the 'grass is always greener' option. US equities and emerging market futures (Brazil, China, Korea, India). I realize access to these foreign markets is difficult and the payoff is uncertain but I think there's a chance a smart guy with lots of time should be able to get something done there. I do know an individual guy using IB and another broker who does extremely well with US equities. But he's also very very good with computers - and he started around 2006. And by extremely well I mean that $1M days are not uncommon. So I guess it might still be possible, but my sample size is 1.
It kind of makes the whole trading thing a downer for people who didn't happen to ride the wave at the right time. In a more perfect world I believe there should be a pathway from retail auto trading to institutional auto trading. I think it would facilitate information flow and make the industry possibly better understood and more transparent. In early 2000s, this kind of existed and there was also reasonable transition from pit trading to autotrading - though many of those success stories were fairly well connected. The only real option I currently see for new entrants is having the right degree from the right school (that's in the right city).
Also, in my mind - and in an ideal world - people who call themselves trading 'educators' would be looking to open this pathway. Because think about this. Plenty of people spend 30-50K/year tuition for 2 years getting a MFE degree (I confess, I did it too, but in 2001). Maybe 5% of the material is actually relevant to trading - which is the job everyone is trying to get. I am 100% confident that I could take those same kids before they sign up for the program, charge them NOTHING, and in 2-3 months have them developing profitable strategies that everyone makes money on. Instead, educators focus on selling a bunch of TA garbage for thousands of dollars that people can't really afford. I've been thinking about this a while - about how such a program should be structured. Maybe I'll start another thread. Too off-topic.
Jeez. No. These are dice. Just count the possibilities.
But, in the interest of fair play I return with two:
1) After accounting for flicker (however you want) look at all the series of price transitions in the market you trade (over a month) and determine how it differs from one you'd expect given a random walk. That might tell you where to put stops.
2) (yeah, old question). If you meet someone who has two kids and you know that one of them is a girl, what is the probability that the other one is a girl.
Count me in.I have a lot of time and willing to work 26 hrs a day.Even though i`m just getting started,i have a couple of strategies i`m testing now.However, I`m not 100 % confident in them.
lol, count you in? Either a) you do not know how many hours a day has, b) you are way overconfident, or c) have a pretty stale sense of humor. What makes you think you are worth being trained and chosen above others?