Examples
Examples
Try the starting example
- Observed tests JSON [days since study,recall fraction]
- [[0,0.8],[10,0.4],[20,0.2]]
- Forecast day since study (days)
- 30
- Fitted day-zero recall fraction
- 0.8
- Fitted log-decay slope (per day)
- -0.069315
- Raw fitted forecast recall fraction
- 0.1
- Fitted half-life (days)
- 10
Try another scenario
- Observed tests JSON [days since study,recall fraction]
- [[0,0.8],[10,0.4],[20,0.2]]
- Forecast day since study (days)
- 40
- Fitted day-zero recall fraction
- 0.8
- Fitted log-decay slope (per day)
- -0.069315
- Raw fitted forecast recall fraction
- 0.05
- Fitted half-life (days)
- 10
Calculation method
Calculation method
Fitted day-zero recall fraction
A=exp(OLS intercept of log(recall) versus day).
Fitted log-decay slope (per day)
OLS slope on log(recall).
Raw fitted forecast recall fraction
A*exp(slope*forecastDay), with out-of-range values explicitly flagged by UI.
Fitted half-life (days)
-ln(2)/slope when slope<0; else No decreasing half-life.
Symbols and units
- a
- Observed tests JSON [days since study,recall fraction]
- b
- Forecast day since study (days)
Use the units shown. A list keeps the order in which you enter its values. The symbol ^ means “raised to the power of”.
Assumptions & limits
Assumptions & limits
- Descriptive log-linear fit only, not a validated memory law. Does not account for review, testing effects or heterogeneous words; forecast uncertainty is not quantified. Display warning if fitted recall outside [0,1]. Supported limits: 2..100 tests; unique days>=0; recall in (0,1]; forecastDay>=0; finite exponential result.
Are these inputs private?
These calculations run in your browser. A shared link contains the values you choose to share. Only questions submitted to the separate AI finder are sent to OpenAI.
Further reading
Background reading; this does not imply endorsement or independent review of SumFizz.