Heya Brian, Kudos for the ideas, I will focus on just a few since I don't believe it physically possible to respond to what you mentioned as a whole, my wrists would kill me, some comments inline.
Not sure if you've read this whitepaper, but I thought to toss it out here since it incorporates some of what you'd be interested in doing [ignore the admittedly biased viewpoint in the article and just read about what some of the buy-side and sell-side algos do] http://www.themistrading.com/article_files/0000/0348/Toxic_Equity_Trading_on_Wall_Street_12-17-08.pdf Also there is some data floating around on the volume of HF trading as a % value of the market volume as a whole, and the numbers are really quite high, around the 30%+ region. This comes as a direct result of institutional competition, and it is no wonder that it has filtered down to us as well :) > - it isn't practical for me (at this stage) because MarketMarkets play the > spread without any frictional costs and they are sitting on top of the > markets with highspeed gear (compared to myself only) > - it might not ever be practical for the majority of traders to trade > microevents OR trade them via AB .... however, like AB/Tomasz, I want to > gather some evidence from real world testing, to find out for myself, and not > rely on others beliefs about whether it is viable or not. I've had some experience in trying these ideas out on both sides of the fence, so I thought to pitch in my 2c. Given what you read above in the whitepaper and knowing that such technologies do exist and are used everyday, let us consider what goes on in the MM world, where companies exist with superior hardware (fast), direct connections (lower latencies), exchange memberships (lower costs) for vast majority of products. Such entities not only optimize hardware, software, operating systems but also in tweaking the languages with which these programs are written in, seeking to get the fastest possible performance. Pretty much ~any~ such entity can generate trades in low double-digit milliseconds (some in single-digits), and that's counting everything in between from the processing to the sending to the receiving. That's right - no graphs, just stats, since all the action happens literally too fast for the eye to follow, not to mention the brain. On the other side is where the majority of us reside, equipped with our home equipment (ok, maybe you splurge a little and get a high-end server with extra ram, a dedicated internet connection, top of the line router hardware and of course, Amibroker as one of the fastest analysis packages out there :), but you are not on your best day going to get speeds coming close to these. This disadvantage is offset by the fact that as customers many exchanges will enable preferential treatment for us, usually in the form of getting a free pass to the front of the queue, or other such priorities, which you can find if you dig into the rulebooks for your favorite exchanges. Given all this, is it still feasible to do semi-HF on the client end? Yes I think so. Customers get preferential orders and the fact that limit orders cannot exceed your expected price, and you get a rebate for providing rather than taking liquidity, even if your broker hates you for it. So, even though any institution can run circles around your orders, your orders can still possibly get price-improved and I think the effort of getting bid/ask data and playing the limit order game is well worth it be worth it, a la Herman posts on ideas of utilizing bid/ask instead of just hitting the market. Now, would you want to try and play the market-maker role? I wouldn't, but if it works out for you, I'll be first line to buy drinks. > Realistically: > > - can AB achieve this? > - is it desireable for AB to achieve this? > - should AB keep its focus on being a cutting edge BackTesting engine, with > some secondary RT charting that is only intended to supplement design and > testing efforts, and leave RT high frequency trading to other software? > - at the least, shouldn't AB allow us to somehow create an extreme event > database and do our testing within AB ... and do this with off the shelf > features, so that it is accessable to all? > - perhaps thinking about how AB could hyptothetically trade microevents, in > RT, will lead to some realisitic options that will end up being implemented > by AB, for use at subsecond timeframes (not necessarily bid/ask trading). Given what I said above, I still think Ami is the best tool for backtesting and for normal RT trading, and either of the per-chart refresh rate, millisecond-timestamp, or even just getting the bid/ask ideas would help a lot :) I'm not sure its desirable to go much further beyond that since any HF trader in search of speed can ditch graphics first thing to chase after faster processing. And if you're going into the sub-seconds, the human eye isn't going to watch or process anything as fast as the logic you wrote in the first place. Not to mention the high cost of entry into this business, which explains why there is no commercial software to really do HF since anybody with the money to obtain such speeds can afford to write/hire their own code without being locked into a third-party provider. If you get to this level of business, let me know :)
