After about a year of heavy R&D, here are some of my high level thoughts:
A) If the role of a trading firm is broken down by 4 key operational functions: strategist, analyst, trader, and risk manager. I believe in the next 5 years, all except the head risk manager, who really functions as a compliance offer will be replaced by a computer. 80% of the job of the analyst will be replaced and automated. The Trader for sure will be eliminated, and about 20% of the Strategists job will be automated (more or less the overlap with the analyst).
B) Within two years, all serious trader will be assisted by some type of trade sequencing software instead of manually entering in every trade. (I think that ratio is currently at around 40% from the last stat I saw)
I define serious here as a player in the market that needs to move size and volume. Obviously a one-two lot manual trader could still effectively trade with wide stops, but it would be a Herculean task to consistently crack the 7 figure income mark while maintaining under .1% drawdown completely manually. There's a reason why lots of the old floor traders have either retired or become "consultants" instead of trading the markets themselves anymore. It's simply impossible now to put on size effectively without computer assistance to mark the best possible prices.
C) For the time being, an elegant Gamma Scalping strategy can still be designed and hard to emulate with a machine. Although the heavy duty Greek number crunching obviously needs software and automation, there is still a large degree of unstructured data that is hard to automate properly.
I had thought when first starting this thought experiment, that the competition was dualistic, i.e. fully automated machines versus manual discretionary traders. I think the future is more likely down to fully automated machines vs "cyborgs", some discretionary strategic decisions assisted with heavy duty automated execution to express the trade idea. The manual mouse clickers I believe will all fall by the way side.
Anyways, thanks to everyone who has helped along the way, those who have joined the team, those who helped me with the silly questions in the beginning like "what makes a good programmer thread".
Here is the current and final tally for the year:
Manual Gamma Scalping Returns: 38%
Automated Retail Partnership Algo Model: 42%
A sample snap shot for how to computers out performed an aging mouse clicker attached.