The Mana World Forums

o11c wrote:

Freeyorp101 wrote:I can confirm that we had in excess of 200 players online half a decade ago. Player count is declining.


---Freeyorp

Right, but it's the time after that (2-3 years ago) that we're arguing about.

The most I ever saw was about 120 in the summer.

Whenever I log in these days I'm always shocked at how empty Hurnscald is now compared to when I used to play regularly. I don't have any numbers to back that, but I'm 100% sure that there's less people around compared to 2009-2010.

o11c wrote:2-3 years ago) that we're arguing about.

The most I ever saw was about 120 in the summer.

120 was the highest peak this summer, while we had more than 100 online chars every day between 18-24h , 50-60 chars online between 6-10h (server time) almost every day, at least, during some periods in 2010-2011. It was difficult to complete the new quest during 3-4 days after the release because players cleaned the maps,( for example It took me 2-3 days to realize that there were more than a couple of bandits in affenpeter) or because you had to wait for 2 or 3 players to have completed the step (Rossy).

It might be best to also consider the multiple character stacks that counted in the online list at the time as well...sometimes people would stack 10 characters at once to level before the rule of no stacking existed.

regards,
chicka

Chicka-Maria wrote:It might be best to also consider the multiple character stacks that counted in the online list at the time as well...sometimes people would stack 10 characters at once to level before the rule of no stacking existed.

regards,
chicka

You're absolutely right Chicka, but not only stacks, sitting alts too. In another post I give a rough estimate of alts regarding total number of players. Now if we look at the online list number and suppose that the number is a fluctuating number with constant average, the number of chars and number of players vary the same way (just a scale factor towards lower values).

Note that those stacks were not necessarily made of alts of a single player; sometimes they were in fact several players but commanded by a single one. This is why the "no follow rule" was installed. though the botting behaviour was obvious.

Based on some raw numbers I was given:

• There were approximately 10 more players at any given time in 2011 than in 2010
• There were approximately 10 less players at any given time in 2012 than in 2011
• There are approximately 10 less players at any given time in 2013 than in 2012
• There is enough noise in these numbers that +10 -10 -10 does not necessarily equal -10.

I'm not much good with statistics so this has all been rough

So 2011 was kind of the "golden age" of TMW...

o11c wrote:Based on some raw numbers I was given:

• There were approximately 10 more players at any given time in 2011 than in 2010
• There were approximately 10 less players at any given time in 2012 than in 2011
• There are approximately 10 less players at any given time in 2013 than in 2012
• There is enough noise in these numbers that +10 -10 -10 does not necessarily equal -10.

I'm not much good with statistics so this has all been rough

If you are speaking of averages Then 10-10-10 =-10

Nard wrote:If you are speaking of averages Then 10-10-10 =-10

I'm not speaking of averages, I'm speaking of eyeballing.

Because analyzing this data *properly* requires a 5-dimensional graph (tod, dow, doy, y, count), which is kind of hard to visualize, especially with the sparseness.

o11c wrote:

Nard wrote:If you are speaking of averages Then 10-10-10 =-10

I'm not speaking of averages, I'm speaking of eyeballing.

Because analyzing this data *properly* requires a 5-dimensional graph (tod, dow, doy, y, count), which is kind of hard to visualize, especially with the sparseness.

Sorry, i could not tell much more than I did because the "approximately x more" is vague enough to hold many different things. Choosing a correct data representation in order to be able to draw conclusions is sometimes tricky.
btw if you used a graphical representation to conclude and trusted your eye (which is the regular procedure), you refer to some central statistical parameters: an average (arithmetic, should give good results in such distribution), or median (same remark). The eyeball radius is more problematic because the impression you get is strongly tight to the parameter you chose. I also made this remark to freeyorp about the statistics he published some time ago. Usually he balls radius should be chosen proportional to the square root of the count of the class you represent so the ball area is proportional to the count.

now if a variable is transformed linear: X ----> Y = a*X+b then arithmetic mean, median, min, max and quantiles follow the same transform.
Variance (ie quadratic mean of deviations to average) is multiplied by a² so standard deviation is multiplied by |a|.

I hope this will help.

Frost, you WILL be missed!! I always enjoyed out chats!

I started playing TMW many years ago, but.....

I no longer play TMW because of the politics, backbiting, backstabbing, talking about others behind their backs, making fun of others, potty mouths, player disregard by the devs/admin, and MANY other reasons.

I did my share of some of these while I played too, but it was usually in retaliation of provocations.

Saying that, I went to see how many players there were in the years past and found this:
http://web.archive.org/web/200807010000 ... nline.html
"Saved 106 times between July 15, 2007 and January 23, 2013" So, settle that argument quickly!
(I found myself listed several times during those years. At one time we had 10 members of my family playing this game!)

TMW used to be a fun game! Not anymore! I'm thankful for LoF! We aren't big, but the players we have are loyal and know thier input is as important as the devs/admin.

KeeKeeKat/Patricia

KeeKee wrote: Saying that, I went to see how many players there were in the years past and found this:
http://web.archive.org/web/200807010000 ... nline.html

I wrote a script to download them all in a more convenient form, and attached the result.