reference_data.Rmd
Here is the reference data used in the scoring. Feel free to use the buttons to download this info for your own analysis!
Used for calculating daily risk-free rates for excess returns and for calculating beta with respect to three market benchmarks: SP500, SHY, and BTC. These benchmarks were chosen because:
The SP500 is probably the most most commonly used market benchmark in finance and financial reporting
We wanted to include a benchmark that had virtually NO correlation with the SP500 so we selected SHY
BTC-USD is interesting and timely, so it’s interesting to include.
Key:
Date | Date on which the value in a column was observed |
SP500 | Closing value of the S&P 500 index |
SP500_rtn | Daily return of the S&P 500 index |
SHY | Closing value of the iShares 1-3 year treasury bond ETF |
SHY_rtn | Daily return of the iShares 1-3 year treasury bond ETF |
BTC-USD | Closing value of Bitcoin/USD exchange rate |
BTC-USD | Daily return of the Bitcoin/USD exchange rate |
3_mo | Daily 3-month maturity treasury par yield curve rate as reported by the US Department of the Treasury |
3_mo_apy | US Treasury 3-month yields converted to Annual Percentage Yield (APY) using the USDT’s formula |
3_mo_td | 3-month APY expressed on a per-trading-day basis by dividing by 252 |
Why BTC-USD is interesting: If you wind up with a large beta (positive or negative) with respect to BTC-USD, then that means that whatever trading strategy you’re running is “accidentally(?)” closet indexing Bitcoin. Usually closet indexing is a bad thing, but consider: Certainly you based your trading strategy on valid assumptions about the market. If your strategy turns out to have a strong beta with respect to BTC-USD, then that means that whatever market principles are at work in your strategy are also at work in Bitcoin price discovery. That means that just maybe, with some insight, you can take a deep look at your strategy and do something that many consider almost impossible: model the price behavior of Bitcoin reliably!
Calculated as gmrr of excess returns observed up to the given date.
One unique complication: when a trader has an odd number of very negative daily returns, the arithmetic mean rate of return is used because the geometric mean can’t be calculated – you can’t take a real-valued root of a negative number.