fighting bad numbers with worse numbers?

In today's Linden blog* entitled "Second Grade Math", The Yankee Group is lambasted for saying that "that the average time spent per user of Second Life is just 12 minutes per month". Linden Lab refutes that with their own metrics:My response, with extra special snarkiness follows. My question is: have I made an error of interpretation?

*http://blog.secondlife.com/2007/10/05/second-grade-math/

------
So who exactly is in remedial mathematics, Ginsu?

[discussion of dead end analysis elided for brevity]

So let’s check your conclusions for reasonableness. You did a simple, yet logically specious division of player hours / number of logins and find that for each login over an entire month, about 23 hours were spent per login. Most of the humans I know sleep sometimes. They also work, eat, fight with their families, etc. So your average login spends less than an hour sleeping, eating, attending to personal hygiene, and so on. Could this possibly be a realistic number? No, it is a mean, that would indicate that some logins were playing SL for more than 24 hours per day.

Keep practicing and you might not have to go to summer school.

I believe that is 23.6 hours per login for the entire month of August. Not 23.6 hours per day.

-Atashi

But ya gotta love them for trying lol

Actually, I think Ginsu was saying that they spend 23 hours PER HOUR in SL..

ROTFL

23 hours per hour?

Head twists and slowly implodes into itself... now THAT is new math.

/me grins

I'm inclined to think that the "974,000 user logins" really meant 974,000 unique *users* who logged in one or more times during the month, as opposed to 974,000 login *sessions*.

If it were the latter, then to run up 23 million hours of usage, there'd have to be a huge number of sessions running continuously for the whole month, hence counting just one session in the stats, but accounting for about 720 hours usage each. And, with enough bots that could be possible (although updates and outages would force a relog).

(Meade, to quote Dorothy Parker, "Cleveland? Yes, I spent a week there one day."

That may be right, in which case it is still 47 minutes per day per login.

But Ginsu (graduate of Princeton and NYU law) has mixed his units and and winds up with a result in hours/login yet reports hours/user. Even still as the two parts of his ratio are expressed per month, the "month" factors out.

Sometimes someone is so confusingly wrong that it is even hard to figure where the error actually lies.

Averages can be fairly pointless anyways. And when the numbers you are averaging may or may not be accurate or make-believe, it makes it even harder to figure out what's accurate.

As my dad used to say, if you stuck your head in the freezer and your ass in the oven, then on average you'd be quite comfortable.

-Atashi

Don't be silly. Of course he's talking about values for the entire month.

It's totally obvious to me but I'm not one of them "must call LL idiots at least once a day" posters..

LL, stop hiding the real info and there will be no speculations!
Consider the countless Alts and many people that log in a few times a year or never. The number of registered users grows while the other numbers stay about the same or grow insignificantly. I'd say LL speculates with information, not the other way around, they want to look much bigger than they actually are, it's good for the press and will help if they go for sale.

Considering the addictive nature of Second Life, and MMORPGs in general - I think an average playing time of 23.6 hours per month is probably pretty accurate, if understated.

Today, my in-world time is limited to about 4 hours per month, due to RL issues. I still spend a good 10 hours a week dealing with ancillary second-life stuff... I just do it all offline.

When I didn't have these RL issues going on, I was easily spending well over 200 hours a month in Second Life, in-world. That's 6 hours a day once I got home from work, till I went to bed - and 12 hours a day on the weekends.

OMG I love that!!!

I play an average of 10 hours a day

Yeah, but, if the shoe fits....

"If you stuck your head in the freezer and your ass in the oven, then on average you'd be quite comfortable."

This is hilarious, LOVE IT! Does it work the same way if your head is in the oven and the other part is in the freezer?

I can attest that several only logged in for less than 2 minutes.

I suspect that this certainly average is highly skewed. Simply logging in for that 1 second, would count towards total usage, then there are those I know who NEVER log off. They eat, sleep and live in SL, then camp when they sleep in RL!

I'd be inclined to say that if we were to force the minimum login amount to at least 10 minutes, those numbers would be a bit more accurate.

Or even by specific IP. I don't know anyone who doesn't have at least one alt. Therefore, if you do it by individual IP, you'd get a helluva lot less log ons.

Its obviously 23 hours per month.

When you add in the fact that there are 1.4 accounts per actual user you get 32 hours per user per month. Which totally passes the reality check.


Edit - in fact he states Months right in the description - Since his simple calculation never canceled out Months it has to stay.

*sends OP and Ginsu back to Diff EQ*

Some months I average >18 hours per day in SL.
Some months I average <18 minutes for the month.

If there's one thing I'm certain of... there are no 'average' people in SL.

I don't know where you people (nor the Lindens) are coming up with these numbers. Over 9 million users? 974,000 logins? 1.4 accounts per user?

We all know that total user count is way exaggerated since it's a count from day one without any known corrections due to canceled accounts, users quitting because of dissatisfaction or disinterest, additional accounts (alts), accounts who could never log in due to machines incapable of running SL, etc, etc, etc. Imagine if the United States census was taken like that........adding the population every ten years without adjusting for deaths, folks renouncing their citizenship. Hell, we'd be 100's (or even 1000's times larger in population than the entire world!!

974,000 logins over a month could mean the same user loging in due crashes (I know I've done that several times a day), same user loging in as an alt, same user running two avatars concurrently, etc, etc, etc.

1.4 accounts per user seems a little low since someone else just stated that they don't know anyone who has only one account..........who's right?

As someone pointed out the numbers could be a "mean"........or they could be an average. Those two are not always the same (in fact they seldom are).

It's sort of fun playing with numbers and presenting them to suit your case.......our governments do that all the time. But unless we have a very firm and accurate base to work with and some very detailed information on how the statistics are arrived at we are likely being fed some of that horse manure that us mushrooms thrive on.

the 1.4 accounts per user was a Linden Stated number.

In reality its probably higher among current residents because the ones who don't stay likely have less accounts each, than those that do.

That only brings the average usage per month higher of course.

Keep in mind that The Yankee Group gets paid to do these "studies" by other companies. One of its biggest clients is Microsoft, who pays Yankee to say that Linux sucks, Internet Exploror rocks, etc. It's a shil for Microsoft and any other company who wants "research" done with predictable "results."

Sort of like RJ Reynolds paying "an independent research team" to determine that cigarettes aren't addictive.

Just sayin....

Of course it does..........that's my point. It's some of that famous funny math.

Do I consider LL idiots as one poster suggested? Nope. Do I think them disingenuous, intentionally deceptive, borderline libelous? Yep. Do I believe them any more than the Yankee group? That's a toughy.

I did notice this whilst trying to give the most charitable possible interpretation of Ginsu's blog entry:To which the natural response seems: "well, why don't you just ask Yankee how they got their numbers instead of publicly implying that they are spreading 'misinformation'?". For this I give LL poor marks in business savvy, which still isn't claiming them to be idiots. For - in the end - more decision makers will look at a Yankee Group report than some poorly written blog entry.

OMG Atashi, I've just fallen on the floor from laughing at this one......gotta write that down in my great quotes of all time notebook......

This thread reminded me of an old Monty Python sketch....

" Well, I was a lad, we used to work down at Mill, 29 hours a day for tuppence a week every six years and when we got home, my dad would kill us and dance about on our graves singing hallelujia"

Don't know why that came into my head.....