At some point in 2003, Tamoran Dal Nava proposed that Wittenberg posts was akin to the Gross National Product of other countries: by measuring the posts of Wittenberg, we would be able to get an idea of the volume of activity in Talossa.
Today, others have made the same observation independently but little did they know that back in 2003, Tamoran had proposed a mathematical formula to calculate the activity which he brilliantly named “The Talossan Activity Index” or TAI.
His formula was mainly theoretical in nature as there was no way someone could sit down and calculate the TAI manually on a daily basis.
Fortunately, Witthoster Marti-Pair Furxheir felt inspired. He created a page which automatically, in real time, calculated the TAI and the various components of the TAI for the current day so far, and the previous 100 days. What is interesting is that on the first day the TAI was put online, we could already have 100 days of history since past days could be easily calculate live in real time.
Granted, in the week of September 2005 (then in the Republic of Talossa), the TAI was cached and updated every few minutes to reduce the load, but that was just a concession to help the server.
Marti-Pair also added a few more statistics, most in close collaboration with Tamoran and others by himself. A little later, Marti-Pair also created the PTAI or Personal TAI which calculated the personal contribution of each citizen to the TAI.
Without further ado, here is a summary of how the TAI is calculated with the key values (each value is better when high except when indicated otherwise):
T: Number of threads for the calculated days.
This should be low compared to the number of posts
P: Number of posts for the calculated days.
The first post of a thread is counted as a post
C: Number of unique users posting
AR: Activity Ratio: P/T, this measured the raw strength of the activity.
L: Load: P/C, this measured the individual activity
TAI: Talossan Activity Index: AR + L, this measured the global activity
CHG: Change in the TAI since the previous day
%CHG: Change in the TAI since the previous day, in percentage
The TAI was a rather good measure of activity since any fluctuation of the 3 raw numbers (Posts, Threads and Citizens) affected the TAI wildly.
Every new post increased both factors of the TAI but a citizen posting a single new ignore thread would reduce both, like we expected it.
But there was something that bothered me with it and I fixed it by creating the RTAI, or Relative TAI as follows:
DL: Distributed Load: T/L (best when low)
I would like to say that like Tamoran, I used advanced math knowledge to arrive to this formula, but in reality, I just plugged numbers until I found an index I liked.
I do not have notes on who created the last index, either Tamoran or I (or both!), but this one is the Volume, which tended to vary wildly:
Vol: Volume: P * T * C
But the greatest index came later, the WIx, or Wittenberg Index. I do seem to remember that I was the original thinker behind the WIx, but I also remember that Tamoran helped me refine it and that he came up with the acronym.
AVGDL: DL Average of the last 10 days
AVGTAI: TAI Average of the last 10 days
WIx: Wittenberg Index: AVGTAI * AVGDL * 61.04
The interest of the WIx was that it wasn’t calculated just on the current day, but rather, was amortized over the next 10 days so that a very active day out of the ordinary would raise the WIx over the next 10 days.
A range of colors was created for the WIx and the data points were as follows:
- Bright red: under 300
- Red: under 400 (but over 300)
- Orange: under 500 (but over 400)
- Yellow (yellow): under 700 (but over 500)
- Green: under 900 (but over 700)
- Cyan (cyan): under 1000 (but over 900)
- Blue: under 1100 (but over 1100)
- Pastel Blue: over 1100
The same coloring scheme was used for the RTAI, with the levels at 1,2,4,8,16,32 and 64.
Overall, my favorite index was the WIx, but between you and I, I have absolutely no recollection of what that 61.04 was from.
Next month, we will review the PTAI or Personal TAI.