There are lots of reasons, none of which are due to overpricing or odd supply/demand issues. One of them that you might not consider is capacity. A fiber line can carry enough bandwidth to sell to many firms. Microwave capacity is very limited and needs to be allocated relatively sparingly. As another user mentioned, system setup is also tricky, as you have line-of-sight and weather issues.
I don't know much about trading as a whole, and in particular the type of trading that 'independent trader' typically entails (holding periods longer than milliseconds). The only way I know how to 'trade' is to make markets, in which case my edge is that I'm getting paid to provide liquidity. It's not the type of trading that can be done independently very easily. In addition, given my current payout percentage at my firm, the economies of scale wouldn't allow me to net any more money on my own (i.e. with the shared resources I have access to here, I can make a lot more money here than I could on my own - more than enough to make up for splitting PNL with the firm).
You could say that. But that means we don't have any quants or traders in our 'trading' firm . I think a better way to put it is that the research and strategies that I use to trade aren't applicable across longer timeframes, and I wouldn't be able to economically replicate the resources I need to trade shorter timeframes if I went off on my own.
what are the monthly costs your firm is seeing? and what do you think the outlay would be for a new entrant? thanks
You can buy yourself a server for $5K and connect to some exchanges for as little as $5K/month. Our costs run in the neighborhood of hundreds of thousands per day. It all depends on what you want to connect to, how good you want those connections to be, and how many of those connections you want to connect to each other. You can run into 7 figures a day if you're connected to all the major stuff globally and are able to hire a good number of good people to build and maintain your stuff. Rule of thumb is $1K per head per day + hard costs.
This is an interesting topic and I'd like to get past page 3 someday. I respect all the input I've read so far; it's certainly not something I can comment on except to say this: human being make orderly markets. they were run off by the machines, which work much more cheaply. Now you get machine markets. You can spin the finer points anyway you want, and the bottom line is still the same. Humans make orderly markets. Machines can't. In the first place, that's not their mandate. It WAS the mandate of marketmakers and specialists. But we've outgrown them, somebody has decided. On the topic of the short term edge disappearing, I've looked over some of my old posts here and see that it was something I was bemoaning ten years ago. The shorter the time horizon, the more random the price action. It seemed to be exaggerated during the particularly dry markets of early 2002 -- when you couldn't buy volume. Trends develop over an increasing time horizon. The onslaught of algos and HFT didn't create that reality, though that doesn't mean they didn't eat up whatever edge you had, if for no other reason than they don't tip their hands in a way that is as recognizable. Bon chance!
You remember 1987 because it happened once since 1929. It, and the "many other occasions" you cite include automated trading. You really don't want to debate this question, do you? The term "flash crash" is older than you are. It's a modern, machine-generated phenomenon.