I want to be very clear -- I am not asking about your available trading capital or development costs. How much do you think it's going to cost to operate this system? What I mean is, how much do you think, after paying someone to develop the system, what will be the costs of purchasing the necessary hardware as well as the costs to operate the system on a monthly basis? Again, I am not asking about trading capital. In fact, assume 0 trades are placed (so there's no commissions, exchange fees, clearing fees, regulatory fees, etc to deal with) and assume 0 labor costs. Just the pure costs of operating the technology behind this if everything were set up and working and you could just sit back and let it do its thing. What would the base cost be per month? The answer to this question gives great insight in to what level of complexity/sophistication you have in mind without getting in to the details.
@Lee: initial prototype phase through Interactive Brokers, 100 market data lines, practically free. Next phase, add up to 1000 data lines, $300 / month. Then a FIX interface with them is $1500 / month. So $50, $500, $2000 as rough estimates for various phases of the project if it goes well. No idea about exchange colocated servers and all and no point thinking about them now with a capital less than half of what a web developer makes in a year in US.
Got it. You're not looking to use FPGAs to process market data on servers colocated at exchanges for $50k. You're actually looking to build something within reason -- and that has the drawbacks of latency, but I'm assuming your pricing model has sufficient edge to overcome this. If you're just looking at using IB API for data and execution, my first instinct is to wonder if you'd be better off using an off the shelf market making solution until such a time that you wanted to expand in to a more elaborate system. Unless there's something specific you'd like to do that can't be accomplished with an off the shelf solution as phase one, you're in for a lot of unnecessary headache trying to roll your own. There's a lot of edge cases that need to be dealt with that I'm assuming off-the-shelf systems already handle. The problem with these edge cases is that they're not necessarily obvious, common, or easy to test (which of course is why they're called edge cases). Simple examples of special situations that are not commonly encountered, but need to be handled -- trade halts/resumes, bad/erroneous quotes/no quotes for a while implying a hung data feed, connectivity issues to the IB gateway/IB gateway and IB servers, misc errors returned from API calls, among others. Be aware that IB's data is consolidated and slightly delayed. Also 100 data lines (or even 1100 when you purchase 10x booster packs) is not very much when it comes to options unless you intend to be tightly focused. There's 3rd party data providers that will be cheaper than 10x quote booster packs, yet give you better data (with respect to consolidation and latency). May be worth considering early in the process. Another potential problem is your order:execution ratio. If you're making markets, your orders need to be updated whenever your pricing model changes the price (which should include the underlying's price movements, which obviously is changing frequently). The result is that your system will be creating, modifying, and canceling orders at a very high ratio compared to orders that are actually executed. IB does not like it when you have a high order:execution ratio. Also because of your high number of orders, you'll get marked as a professional fairly quickly and deal with the consequences of that. I have the technical skills to bring this to life and I have the other $50k you were looking for. I also understand a lot of the other challenges involved in doing this on a small scale, which is why I've not done it myself. What makes your method special? Sell me on it.
You don't have to develop a "market maker system". There is no "system". If you want to do it properly... what you need, is knowledge about options... and more than you think you need... And you need a trading platform that allows you to mass-quote and very quickly adjust pricing parameters, mainly IV. And it needs to be fast enough, and on fast enough connections. Those "systems" you can just get off-the-shelf. But you need the infrastructure around it as well. You can't just rely on 1 product to trade... you need more, since you want to spread out risks across stocks... across sectors. You could opt for just trading an index, like S&P500, but you're likely missing tons of info regarding pricing because you don't have eyes on other indices. Indices are the most efficient markets, so there's hardly any juice... unless you hold larger vega positions... for which you need a whole lot more than just 50k. In the end... 50k in true market making doesn't get you anywhere. You need way more capital to be able to hold hedged positions. 500k doesn't even cut it anymore... I suggest you keep that 50k in your pocket and trade little bits in options in the next 1-3 years to learn about the dynamics of options. Learn about probabilities. Learn about risk. Learn about hedging... dividend, interest rates, time-effects... @Lee- has done an excellent post. With IB you will run into massive issues regarding mass-quoting/orders. You will also run into "heartbeat" problems... frozen data. Especially when you will need it the most, in heavy volume and moves... you will not be able to use IB as a reliable source for quoting purposes.
@Lee: >> There's a lot of edge cases that need to be dealt with that I'm assuming off-the-shelf systems already handle. And problems such as machine power down during bulk send of orders and trade reconciliation. I'm definitely not gonna write a spaceship from the first try. For starters I'm aiming for "good enough". >> If you're just looking at using IB API for data and execution, my first instinct is to wonder if you'd be better off using an off the shelf market making solution until such a time that you wanted to expand in to a more elaborate system. Can you recommend one or two such systems? So they have to work with IB API. I'll check how much they cost, what can they do and what customization do I have to add in order to use my pricing models and strategies. >> Unless there's something specific you'd like to do that can't be accomplished with an off the shelf solution as phase one, you're in for a lot of unnecessary headache trying to roll your own. >> I have the technical skills to bring this to life and I have the other $50k you were looking for. I also understand a lot of the other challenges involved in doing this on a small scale, which is why I've not done it myself. What makes your method special? Sell me on it. First, I've my own models and strategies which I need to use / integrate with another software. I suspect the learning curve and hassle integrating with a closed source third party is comparable to rolling my own, plus I can't customize it and I don't own it and have to pay for it. As opposed to possibly selling mine to others (I've also got some ideas about also selling what I'll develop but I'm not discussing them here). I'll just implement a prototype by myself and run it for a while.
>> @JackRab: I suggest you keep that 50k in your pocket and trade little bits in options in the next 1-3 years to learn about the dynamics of options. Learn about probabilities. Learn about risk. Learn about hedging... dividend, interest rates, time-effects... I know this stuff man, I've published articles with my own option pricing models (those that don't work, of course