I know statistics/inference/timeseries, but they are only on paper. Being able to deploy these techniques (data science?) to vast intraday data is something I wanna do.
Ok, understood. Below maybe a good place to start using python. https://info.cloudquant.com/products/#freelancer
i mean, am overwhelmed by the idea of starting even a toy trading system, for someone self-taught. There are so many components to build, from datafeed handlers and signal generators to trade engines and risk mgmt etc. Too many parts and none trivial (data-intensive, threads, socket programming etc), don't even know where to get started. Prefer something simpler to toy around first.
Why not starting small? - Get historical data like from quantquote.com or so. - Write python code that implements for trading strategy and backtest it. I would not implement trading system before you have working trading strategy. You don't need thread or sockets for trading system. Real-time data feed you will get from your broker like IB, TradeStation
My trading will be intraday-only based on 1-min prices and multivariate time series, so presumably will use many callable/future threads for calculation.
if you’re just using price series, something like NinjaTrader or tradestation is probably fine to get started and program some simple strategies. You can always roll your own later.
sp 79, I do have an interesting project that needs collaboration - predicting major market reversals. If you`d like to see more, ( charts/ results) pm me.
True. Without spread it would be possible to get enormous numbers out of the market, with almost linear growth without noticable drawdown. But if considering more factors it becomes much more challenging.