QC500 and SP500 historical constituents lead to dramatically different results.

Below are the results of sp500 historical constituents

I believe I found the problem. Unfortunately, the issue is with the good performer rather than the bad performer. 
When using historical constituents, I  was incurring in selection bias. I was picking a lot of companies with high growth but before they were added to the sp500. So those stocks were the ones giving me the best returns. 

I am still looking to see if this strategy can be optimized or dumped altogether :) 

Hi DarioA
lthough it is set to mimic SP500, QC500 selection is not necessary to be the same as SP500 constituents. They are selected in a daily basis based on dollar volume. Please refer to this source file on its logic.To select SP500 constituents, we recommend using the ETF constituents dataset with SPY.
Best,
Varad Kabade

Is it expected that for some days there are significantly less than 500 constituents in the SPY? Here are the top 25 dates with <500 constituents. The data was obtained using the `ETFConstituentsFilter` given here.

Date          n_spy_constituents
2016-01-15     16
2016-01-19     16
2021-12-01     65
2019-01-09    111
2018-01-29    152
2017-11-13    243
2021-12-23    250
2022-03-31    254
2015-02-25    311
2016-08-24    315
2018-12-10    483
2018-10-02    493
2018-12-26    494
2022-07-18    495
2010-03-23    496
2010-03-17    496
2010-03-19    496
2010-03-22    496
2010-03-18    496
2010-03-24    496
2011-04-28    496
2010-03-26    496
2010-03-29    496
2010-03-30    496
2010-03-15    496

Hi Cameron,

Thanks for reporting this issue. To track our progress of resolving it, follow this data issue.

Best,
Derek Melchin

Want to invest in QuantConnect as we build the Linux of quant finance? Check out our Wefunder campaign to join the revolution.