Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
You guys are dangerous.

iceman.jpg
Snort!

I feel the need for speed! :D

Chart #3: "End of Time" Top 50 Posters

Computed by Doctor Q in July, 2009
Dr. Q, you've probably mentioned this already, but what number/date do you use for your "End of Time" period in Excel? Curious.

Thanks again for all the data analysis. A very enjoyable read.
 
It's true. My apologies. It was a bit rude to say that Tallest Skil's predictions were more wrong than Q, and it shouldn't have been said, even as a joke.

Sorry Q.





Oh, and sorry TS.
;)
 
Dr. Q, you've probably mentioned this already, but what number/date do you use for your "End of Time" period in Excel? Curious.
I don't use a date at all. I simply put people in order of their post rate, although I use tricks to pick an appropriate post rate for each person based on both long-term and short-term posting.

Here's a way to see why I don't need to pick a specific date for the End of Time:

Suppose I have $100 in the bank and I deposit $5 every week.
Suppose you have only $50 in the bank but you deposit $8 every week.
Suppose Mrs. Q has a whopping $150 in the bank but deposits only $2 every week.

Who will be richest at the End of Time?
Week 0: Me $100, You $50, Mrs. Q $150
Week 1: Me $105, You $58, Mrs. Q $152
Week 2: Me $110, You $66, Mrs. Q $154
Week 3: Me $115, You $74, Mrs. Q $156
Week 4: Me $120, You $82, Mrs. Q $158
Week 5: Me $125, You $90, Mrs. Q $160
Week 6: Me $130, You $98, Mrs. Q $162
Week 7: Me $135, You $106, Mrs. Q $164
Week 8: Me $140, You $114, Mrs. Q $166
Week 9: Me $145, You $122, Mrs. Q $168
Week 10: Me $150, You $130, Mrs. Q $170
Week 11: Me $155, You $138, Mrs. Q $172
Week 12: Me $160, You $146, Mrs. Q $174
Week 13: Me $165, You $154, Mrs. Q $176
Week 14: Me $170, You $162, Mrs. Q $178
Week 15: Me $175, You $170, Mrs. Q $180
Week 16: Me $180, You $178, Mrs. Q $182
Week 17: Me $185, You $186, Mrs. Q $184
Week 18: Me $190, You $194, Mrs. Q $186
...​
As you can see, at the start the person with the most cash (me) is richest. But if you wait long enough, the person who deposits more (you) will pass anyone who deposits less. The only effect of your initial cash is to determine how long it takes before you win or you are left in the dust.

Therefore, to answer the question of who will be richest at the end of time, I don't have to go through week by week. I just look at who deposits the most per week and put everyone in order: You'll be richest, I'll be second, and poor Mrs. Q will be poorest.

In the same way, the person with the most previously made posts gets a high ranking now. But the person who posts more often will pass anyone who posts less often, if you wait long enough.

Example: Suppose a member named Hamachi has 11,225 posts and posts at a rate of 4.41 posts per day. Only the 4.41 determines his rank at the end of time. If you wait long enough, he will be ahead of anybody who posts less than 4.41 times per day and behind anyone who posts more than 4.41 times per day.
 
im guessing Dec 12, 2012. :p
Snort! Good one! :)

I don't use a date at all. I simply put people in order of their post rate, although I use tricks to pick an appropriate post rate for each person based on both long-term and short-term posting. <snip some good raw number stuff. :) >
Thanks for the explanation. :) I think I understand your explanation as it makes sense to me.

However, how far out do you need to go until you are sure the ordering will not change?

For example, using your sample data, how do you know not to end at week 10 and go until week 18 to determine the end of time figures. Do you go out until there is no change in the ordering once you apply your specialized post rate? Or do you have an Algebraic or other mathematical approach to the problem?
 
~snip~

Here's a way to see why I don't need to pick a specific date for the End of Time:

Suppose I have $100 in the bank and I deposit $5 every week.
Suppose you have only $50 in the bank but you deposit $8 every week.
Suppose Mrs. Q has a whopping $150 in the bank but deposits only $2 every week.

Who will be richest at the End of Time?

~snip~

As you can see, at the start the person with the most cash (me) is richest. But if you wait long enough, the person who deposits more (you) will pass anyone who deposits less. The only effect of your initial cash is to determine how long it takes before you win or you are left in the dust.

Therefore, to answer the question of who will be richest at the end of time, I don't have to go through week by week. I just look at who deposits the most per week and put everyone in order: You'll be richest, I'll be second, and poor Mrs. Q will be poorest.

In the same way, the person with the most previously made posts gets a high ranking now. But the person who posts more often will pass anyone who posts less often, if you wait long enough.

Example: Suppose a member named Hamachi has 11,225 posts and posts at a rate of 4.41 posts per day. Only the 4.41 determines his rank at the end of time. If you wait long enough, he will be ahead of anybody who posts less than 4.41 times per day and behind anyone who posts more than 4.41 times per day.


Yah... and it shoulda worked for Lehman Bros too. But I guess they forgot to post while they waited for the end of time!
 
is there a way to see your posts per day in a more recent since? cause i hardly used my account within the first year i had it, so its thrown off my PPD count, but have been posting a lot more this last year and esp last month. i wish there was a way to see that. cause im pretty sure 1.85 posts per day isn't correct.;)
 
Thanks for the explanation. :) I think I understand your explanation as it makes sense to me.

However, how far out do you need to go until you are sure the ordering will not change?

For example, using your sample data, how do you know not to end at week 10 and go until week 18 to determine the end of time figures. Do you go out until there is no change in the ordering once you apply your specialized post rate? Or do you have an Algebraic or other mathematical approach to the problem?
I don't go week by week (or day by day) at all. It's a mathematical fact that it will happen eventually, so I don't have to check.

Here's a way to see that. If you go far enough to the right on this graph, which line will end up being permanently higher, the red line or the blue line? You can't see where it will happen, but you know it will if you extend the graph to the right far enough. You don't even have to figure out how far you have to extend the graph to make it happen, since you just want to know that it WILL happen. If I had 50 lines to rank, to determine which would end up highest at the end of time, I'd only look at their slopes, not how high on the graph they are.

line_slope.png

Q, out of interest, when do your figures predict Mad Jew will fall from #1 in post count?
I hope I'm not appearing to contradict what I just told sushi to say that I can if I want compute when one particular user will pass another. At the post rates I used for the projections, Tallest Skil will pass mad jew on October 20, 2010. No other user (with more starting posts) will have caught up with mad jew by then, so Tallest Skil will be #1 on that date if everything continues as it has been.

is there a way to see your posts per day in a more recent since? cause i hardly used my account within the first year i had it, so its thrown off my PPD count, but have been posting a lot more this last year and esp last month. i wish there was a way to see that. cause im pretty sure 1.85 posts per day isn't correct.;)
You could search your old posts to find out when you consider yourself "active", then divide the posts you made since then by the number of days since then, but here's a sneakier formula you can use:

Pick a number D that you think represents the number of days you were a lurker before you started being an active poster. If it's only approximate, you'll still get an approximate answer. Then compute this value:
RECENT_RATE = POST_COUNT / ( POST_COUNT / OVERALL_POST_RATE - D)
For example, your profile page shows that you have made 1002 posts and have a post rate of 1.86 posts/day. Suppose you weren't active for the first year, i.e., D=365. You get
RECENT_RATE = 1002 / ( 1002 / 1.86 - 365) = 5.79​
as your estimated recent post rate.
 
Doctor Q, I think this comic fits this thread.
 

Attachments

  • extrapolating.png
    extrapolating.png
    21.2 KB · Views: 160
Cool list, but reminds overly of that saying about statistics:
Statistics are just numbers looking for an argument.

I demand that the Dr. add an extra place for measuring those who post actively but have little to no substance to their membership. Then at least I'd show up on the list. :D
 
I think it should be noted that me and iJohnHenry did whatever we did and had some alone time. ;)

^5 fellow badboy.
 
Thanks Q! It's embarrassing how much I look forward to this twice a year. It's also embarrassing how I've let my posting slip. I think I need to work harder to skew your numbers.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.