Nitro: "I never suggested you are full of shit - only that you are worrying about things that you could not possibly take advantage of, and that the people that can you aren't helping them because they know what the real issues are. nitro" I am the people who can take advantage of the low latencies. I have for years and years. The reason I have survived is because I have eat, breathed and slept these issues. The end result has been survival and profitability. I also know the end result of loosing speed for all of my algos: It equals less to money and failed trading because I am late. Whether it be marketmaking, directional decision making, everytime I take step back in latency by an amount smaller than most people see, I see the end results. I am at the point where I know roughly how much more moeny I make with every certain "increment" of latency reduced. That knowledge has been pretty solid for almost 6 years now. With regards to IB: Nice guys, but they have always been too slow for my needs. Mike Domka has always been a champ and a very nice person to work with exploring if we could have worked together. So far they have been too slow. Now, that doesn't mean they are slow for other high speed guys, just for me based on their stats and my needs. Wow, just got back. Thank god I got out of New York early to drive home with all of that mess. I have alot to read and catch up on here.
"The reward to the programmer for "fixing" any minimal latency at the brokerage is usually NOTHING. Most people won't notice any changes for the better. A few customers might notice an improvement, but these will be the same anal twits that will continue to complain about any remaining latency until there is ZERO latency whatever, which of course is impossible. So, basically they are never happy. Fast is never fast enough for them. The author of this thread is a clear example of this." Actually, the reason for my frustration is that in the 5 brokers we have used, it is the same pattern: They all claim they are faster and better and we come in and prove them wrong. The return curve ALWAYS dips for us with the first 3-4 months at a new broker due to this crap. After much bitching and showing actualy examples, it always (or usually) improves back to the old levels. This pattern is just getting tiring over and over. It's not and never has been an issue with "never being satisfied". It's when you take a huge step backwards and you feel it in the pocketbook. Then, you prove the slowness, fix it and are on track again. Same story and it's all real chief.
I still don't understand.... why don't you set up all the equipment yourself and just have a clearing relationship with a broker? what am i missing here? Thanks, - The New Guy
I have checked with Mike Domka; unfortunately, he does not recall giving out our speed statistics to anyone. Would you mind PM-ing me the data that you have about us and how it compares with the best broker you found? I would really like to understand how you concluded that we are too slow for your needs. Thanks in advance.
Put criterias so that I can understand what you are talking about. Such as: Latency to: client to BD BD back office, buying power checks, shorts, connections (to and from destionations), ADP vs. in house drops to NSCC. Databases, oracle vs. sybase. Unix vs. windows. Exchanges-their network speed(they are all different), acks-true vs. fast. Quotes- where you gettn quotes from or your bd getting them from. Telecom connections- carriers and type of connections. It all seems pretty simple to me. Cash is king in all the above.
Cash is not always king; Knowledge is. I remember 1 year ago we had a more than stellar month. Above average. So what we decided to do was experiment the next month (at the expense of profit) as an investment for the future. We shut off our fastest servers and forced ourselves to run the main algo on an old Dell dual 500 mgz server. Why you may add? To push the limits of the software side. At the end of that test, we matched the performance of the faster servers with a margin of 5-7% less (meaning percentage slower than the newer servers). So basically we created an edge by making hardware secondary. It forces you to focus on one side of an issue. So no, my POS dual 500 mgz server was maybe worth $300 and it won. The reason I tell these stories is because this is and has been my life for a long time. So many replies are testament to the mindset in the community, thus the mindset of the brokers.
Also, I know I have said alot of this forum. I am not trying to "show off" or say i'm better than anyone else. There are many many algos out there that make more money than us however they do it. I just want to prove that as much as brokers say they cater to "high speed" clients, the reality is that they don't know how to, or atleast they are NOT WILLING to put in the extra effort to understand and do so. I felt that coming out like this was really a different approach and maybe a broker or two would wake up and realize they are running like they have no clue (and I want to believe they are smarter than that!). All brokers are good at something. I know I make them all sound lame and that is not the case. They make money on volume because people who trade with them are repeat customers because they are successful. I just get frustrated because I am true high speed and they don't seem to care.
I have spent sometime on both sides, both buy-side and sell-side, and I run a set of automated trading systems (high frequency). The fact of the matter is that you are only as important as the level of commission / financing revenue you generate them. I remember if Pequot even sneezes, the ibank I was with would invest millions (strike that, tens of millions) in buying Kleenex for them. With respect to high frequency trading systems, it becomes a double edged sword, because of high fixed cost associated with establishing such a platform. And whether it is truely the broker's focus (cut through all the sales crap, "every customer is important, blah , blah"), unfortunately even for prime brokerage, the majority of the revenue would come from financing the positions. Almost all high end high frequency firms / traders have their own infrastructure, and they invest easily millions into their own platforms, and would only use the brokers for trade processing and financing, so these firms would negotiate for the bare minimum rates, and barely have any reliance on the broker / clearing firms for technology. The smaller firms would want to piggy back off of the broker's platform. For instance, at a certain ibank, prime brokerage infrastructure is completely seperate from their own internal infrastructure for proprietary activities, and naturally they would pour more money into their proprietary platforms, as it generates higher RR. The client RMDS, while very good, has definite measurable latency differences from the proprietary trading RMDS ticker plant. Sure, the broker's sales people (okay, "relationship managers") would say all kinds of things ("best of breed", blah, blah). But the fact of the matter is that the transactional revenue per unit (per share, per contract) is low as is, most trade processing buisness units (security services) of ibanks have negative P&Ls, and they are only there for flow, the money is in prop trading, financing (which can be done for clients that have their own infra), and structured products. As for myself, I am slowly tearing my platform away from my clearing firm (I am a professional with my own exchange memberships), my market data environment (redundant nxT1s, etc) is already completely seperate. I am planning on completely seperate the order routing from them by Q1, and I am doing parallel testing now. When I left the ibank, I had similar thoughts, "why don't I create a broker specifically for automated trading systems?", I would have cutting edge technology, great developers (and I know more than a few of them), etc. But you know what? At the end of the day, the rate of return for this business doesn't make sense, since the "NewCo" would still have to compete with the regular firms for "mils" on the transactional revenue. Instead, I rather be trading, heh. I truly believe that for a serious automated trading firm, it pays to have your infrastructure under your control, completely.
What I don't get is: 1. Brokers all fight for business. Why not use the guys like "me" as a benchmark to then advertise they really kick that kind of ass with benchmarks to prove it. Even if they could care less about me and my systems. As I stated earlier: They would get more business, volume and commissions. Good post by the way..........
Because, a large organization, somewhat like a dysfunctional family, each party have their own interests. The security services (non financing) business tend to be poor cousin that doesn't get much budget or resource allocation. Yes, the security service sales people would definitely fight over each other for business, because they want to make their quota and keep their jobs. Every business case, for new equipment, better / more developers, etc, need to be evaluated for payback periods, and better latency to support existing "smallish" professional clients can be described as "revenue retention", and would be less important than say "new business" or "revenue growth". I remember a quote that the head of the international prime brokerage business once told us in a meeting, "frankly, I don't give a f*ck about any fund that is less than $100M bucks, or won't be $100M in a year". I was frankly stunned, but that's the reality. it is similar to the Merrill regional private client managing director (one that was quoted on WSJ) saying "well, people with less than $350k I could care less about, they are just poor". it certainly isn't customer servicing, but that's the truth.