Timeline for Reputation's sliding window?
Current License: CC BY-SA 3.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 5, 2014 at 1:21 | comment | added | Włodzimierz Holsztyński | Thank you @JosephO'Rourke for correcting my orthographic errors. | |
| Dec 5, 2014 at 1:01 | history | edited | Joseph O'Rourke | CC BY-SA 3.0 |
weak => week throughout.
|
| Dec 4, 2014 at 10:43 | vote | accept | Włodzimierz Holsztyński | ||
| Dec 4, 2014 at 9:05 | comment | added | Włodzimierz Holsztyński | I used in the past and am fond of weighted averages but here I would not exaggerate. I prefer simplicity and the directness (of the meaning of the periodic scores). This would be my preference. Anyway, a high score will vanish in the $7d$-window but will last longer in the $30d$-window, etc. I wanted things smoother, not smooth :-) | |
| Dec 4, 2014 at 8:03 | comment | added | Joonas Ilmavirta | If the scores should vary smoothly, I find a sliding time window too abrupt. If I wrote a great answer and got a lot of points on the same day, my sliding window score will suddenly fall when the window passes. A sudden rise on the day of answering sounds perfectly fine but the sudden fall is less desirable. I would prefer a weighted sum of daily points over the last $n$ days where the weight goes nicely to zero as day goes to $\text{today}-n$. Perhaps exponential decay with different decay parameters would give a nice collection of slowly varying scores with different time scales. | |
| Dec 4, 2014 at 7:17 | comment | added | Włodzimierz Holsztyński | Thank you, @quid, for adding the reputation tag. Indeed! | |
| Dec 4, 2014 at 7:05 | answer | added | user9072 | timeline score: 6 | |
| Dec 4, 2014 at 6:59 | history | edited | user9072 |
edited tags
|
|
| Dec 4, 2014 at 3:01 | history | edited | Włodzimierz Holsztyński | CC BY-SA 3.0 |
consistent numbering
|
| Dec 4, 2014 at 2:49 | history | asked | Włodzimierz Holsztyński | CC BY-SA 3.0 |