Q3D: Your problem is you still haven't figured out what the markets are about and how they actually work. You're approaching it in a "the markets should work for me and my ideas, damnit!" fashion rather than "I will trade what the market is actually telling me and I will take note of where potentially exposed traders are when applicable." HFTs are annoying, definitely, but by no means do they make trading impossible.
I am sure 99% of the members of ET never invested the time nor the money I invested in learning to understand the market and trying to build a performing system. People like Handle123 understand perfectly what I mean because they did the same: hard work for many years and constant studying the markets. Only then you can be successful, not in 1 or 2 years. High capitalized: if you have a good system and a good understanding of math you can do this without being high capitalized. All depends of you returns, and your drawdowns. If they are very performant you don’t need a lot of money. But I will not go deeper in this. Highly experienced: if you are high experienced or not depends of yourself. If you make an effort you can be high experienced too. For most people it is enough to be not so lazy to comply to this. Discretionary risk management: I use simple logic. If you have a good system, the system will take care of you, and you only need an emergency stop. You don't have to break your head on it. It is better to break your head on the development of your system because that’s what pays of. Outliner and not relevant: not true. I did what logically everybody should do who wants to trade. So if everybody would do what they have to do they would or stop because not successful, or be like me. The losers would then be the outliners and not relevant. PA teachers: my general rule is: what is public or teached to others can only benefit the teachers. Valid things will in general never be teached, even not for a hew thousand bucks. Because it would generated more money in 1 year than all the subscriptions can bring. If a system that was teached would be really successful everybody would use it, which is not the case with any system, and it would stop working then probably. PA teachers avoid many issues because they don’t understand them and cannot master them. So it is lack of their knowledge. But why should they do an effort to understand theses issues? They make money by TEACHING not TRADING. And as long as their clients make no issue from it they are save. I make money from trading so understand and mastering these issues is essential for me to survive. I am not super intelligent, have no MBA or PHD, but I worked hard and long. And what is very important too: I had a very original idea (thinking out of the box) that resulted in an edge that works already over ten years now. Over 10 years means it is a really very good edge. Time has proven this.
This is probably the only quote I have come across so far that validates what I felt when I was initially designing the architecture of my trading system. Several times I ended up questioning myself if spending quality time in getting the design right is more important than developing something quickly that can make money... because profit is the key. I decided to architect it properly... I had to stay away from the market for months at a stretch watching at the volatility helplessly... I guess its a journey one has to go through to in order to have the right kind of trading minion
You can't incorporate intelligence into a system that has to respond under 1-2us... which is pretty much the realm of HFT. A simple division takes a minimum of 20-30ns (approx/avg) and so you must very methodical/through about the operations you want to use to make your system respond within 1-2us. Intelligence involves a lot of computation and its practically not possible to shove all the logic within 1-2us. However, I do agree that if you raise the base to a system that responds under 1ms (which is not HFT from a practical sense), you can have intelligence to some extant. Again, all this depends on ones own definition of constitutes intelligence and HFT. Believe it or not, I personally know a few "profitable" manual intraday swing traders who consistently say that HFT algos can be a good thing if you know how to come up with strategies that suite modern day markets; thats what got me into trading in the first place
Extremely CPU specific and besides the real hardcore pros will manufacture their own ASICs or equivalent.
You can use point specifics ASICs like our Bitcoin folks and get to 100x faster than intel xeones, but the problem with that approach is that any minor tweak (say change a + to a -) in the algorithm will cost millions of dollars + months of wait time... the algos will not be adaptable which is contrary to hft requirements. Not trying to start a debate here, but simple math that engineers use to do rough calculations: a 4GHz processor (single core/thread) has a clock cycle of about 0.2ns and a division takes roughly about 89 clock cycles (~18ns) on a 64bit intel xeon processor (https://www.quora.com/Whats-the-computational-cost-of-the-division-operation), FPGAs performance is similar. There are some optimization techniques, but you can only save a few nanos. The other big issue is memory hierarchy... you can use parallelism, but that has other issues such as cache coherency which makes things even worse. To put things simple, there only so much computation you can do within 1 us. Highly intelligent sw engineers can squeeze more code into that 1us compared to others, but there are limits to computation.