Hi All!

We're excited to share that coming soon there will be L1 quote data automatically included in backtests for US equity stocks. We are also updating the historical and live data feed to be based on the SIP; the aggregated tick feed from all the exchanges.

We strive to make our backtests as close as possible to live trading, but this modeling is difficult without the full volume and a realistic spread. With the new data sources, it should be a significant advance in the realism of the backtests. The new source would be identical to paper trading, allowing your live paper trading to be more realistic than ever.

If you are interested in the engineering please feel free to subscribe to follow along. Once the code is ready and data in place we'll merge the pull-request and hot-swap the data over. There should be no downtime. I will also comment on this discussion thread once the merge is complete.

Important Note #1: Because the backtests will now have spread; your backtest performance will likely be worse than it was before! I understand this will be upsetting but please see it as a more realistic reflection of the model. It means if you buy and sell in the same second, it will constantly lose the spread between the equity bid and ask. This cost will be significant if you are trading quickly or on low liquidity assets.

Important Note #2: Quote data is approximately 3x the amount of data. Although the processing speed of LEAN will get faster in "data points per second" -- the absolute wall-clock time for the backtest will increase. We've got 2-3 massive engineering leaps up our sleeves which should offset that delay by running 300% faster... there will be 1-2 weeks where the backtest will be slower while we install those new engineering advances. 

We expect this upgrade to happen in the next week. Thank you for supporting QuantConnect! Here's to an awesome 2020! 

Best Regards,