Thursday, January 9, 2014

Outtakes in Blogging - New Years Stats Edition

Sometimes, trying to write a blog post ends up with hilariously poor results.  As an outtake, these are amusing to witness from the outside, of course, but there's a lot of frustration involved, as you might imagine.  Still, it's something that -I- can laugh at when presented with it, so it's not something I'm adverse to sharing and that's exactly what I'm going to do tonight.  You see, I -was- trying to figure out a very, very basic thing for an opening stat and, well, it kind of broke me.  I was essentially just trying to work with something without having enough data for it, which I realize now (and honestly realized at a few points, but damnit, I was going to try) and it just did not end well, as you can imagine.  But I'm welcoming you to enjoy my hilarious missteps in the process of trying to provide content because it in itself is content of sorts.  I'm inserting a page break, however, because it's a chatlog, which means it's quite long by nature without being -full-, as paragraphs are.  So, yes, hopefully you will enjoy it in lieu of actual numbers tonight.

good lord, did I wobble this one spectacularly

[00:01] -->| Haplo has joined the chat.
[00:01] Haplo 'Lo
[00:01] Mogs HELLO
[00:04] Haplo ... What? :|
[00:04] Mogs I'm trying my hand at the Year in Review stat post
[00:05] Haplo Ahhh
[00:05] Mogs And I'm trying to project the YoY increase in delay if I keep at it the way I have been
[00:05] Mogs First post was on December 31st, 2011 or Day Zero
[00:05] Mogs Second was January 2nd, 2013, or a Two Day increase
[00:06] Mogs Today's will be January 9th, 2014, or a seven day increase on the two day increase for a nine day increase total
[00:06] Mogs Thus I'm trying to use those figures to project a next year
[00:06] Mogs and I'm getting 14 because of a cumulative growth
[00:06] Mogs because 0->2 = 2, 2->7 = 5, 2+5 = 7, 7+7 = 14
[00:08] Mogs And if that's true, I'm not sure if I then do another full cumulative of 2+5+7=14+7=21
[00:08] Mogs Or just 5+7=12
[00:09] Mogs To project the -next- next one
[00:10] Haplo nodnodnods.
[00:10] Mogs Or, I guess it would be
[00:10] Mogs 5+7=12+7=19
[00:10] Haplo My advice
[00:10] Haplo Calculate the average posts vs days of your last year
[00:10] Mogs "stahp"?
[00:10] Haplo Then calculate your average posts vs days of your current year
[00:11] Mogs How would that correlate?
[00:13] Haplo Oh, you're trying to find the YoY increase in -delay- and not just general increase?
[00:13] Mogs Right.
[00:13] Mogs There's been a YoY -decrease- in general since I've posted less each consecutive year
[00:15] Haplo Ok, then
[00:15] Haplo What you do
[00:15] Haplo Is you do what I said
[00:15] Haplo Then divide your current year average by your last year average.
[00:17] Mogs How is that going to tell me what day I would post the New Years Stats next year if I continue to be as cumulatively lazy next year as I have this one
[00:17] Mogs i.e. the delay
[00:18] Haplo Oh, right
[00:18] Haplo Pardon, I see what you mean now
[00:19] Haplo Alright
[00:19] Haplo Go through every post last year
[00:19] Haplo And note down the delay in minutes
[00:19] Haplo Or days if each post is delayed by a day
[00:19] Mogs haplo
[00:19] Mogs stahp
[00:19] Haplo What's the matter? :|
[00:19] Mogs staaaaahhhhhp
[00:19] Haplo :|
[00:19] Haplo wins.
[00:20] Mogs you're convoluting my convoluted situation
[00:21] Mogs anyway, 14 is the correct answer for this go around at least
[00:21] Mogs meaning next year would then be the 23rd
[00:24] Mogs It's figuring out -that- cumulative growth that's the problem then
[00:25] Mogs Because 0->2 = 2, 2->7 = 5, 7->14 = 7
[00:25] Mogs Which still doesn't seem right
[00:26] Haplo What if
[00:27] Haplo You added up the delay in days between every New Years Stats post you've done
[00:27] Haplo And found the average
[00:27] Mogs haplo, I will fucking rip your throat out
[00:27] Haplo :|
[00:27] Haplo I'm being serious this time!
[00:28] Mogs I only have two delays to work with
[00:28] Mogs Because I have Day Zero, I have the 2nd and the 9th.
[00:29] Mogs Which is an average of 5.5
[00:30] Haplo Ok
[00:30] Haplo So your first delay was the 2012 post?
[00:30] Mogs Yes.
[00:31] Haplo And what was the delay there?
[00:31] Mogs 2 days.
[00:31] Haplo And they 2013 Post was a delay of 9 days.
[00:31] Mogs Right.
[00:31] Haplo Then isn't the delay drift rate 7 days?
[00:32] Mogs That's the number I got with the Cumulative, yes.
[00:32] Haplo Right.
[00:33] Haplo So there you go then.
[00:35] Mogs so wait, would
[00:35] Mogs that mean that next year would be the 16th instead?
[00:36] Haplo At this projected rate, yes.
[00:36] Mogs did you seriously just ruin the last hour of my math with that
[00:36] Haplo That's the 'delay growth'
[00:36] Haplo Yep.
[00:37] Mogs but what about the next year
[00:37] Haplo Not enough data.
[00:37] Mogs yes there is, because you can extrapolate
[00:37] Mogs that is what I've been doing
[00:37] Mogs is extrapolating
[00:37] Mogs extrapolating the shit out of it
[00:38] Haplo You've been extrapolating with...
[00:38] Haplo Two data points.
[00:38] Mogs yes
[00:38] Haplo There isn't a respectable journal on Earth that would buy that.
[00:38] Haplo At -most- your extrapolations for the coming years would be extremely loose.
[00:38] Haplo And would slowly change with each year.
[00:40] Haplo Because you -could- say that due to this growth rate, the 2014 Stats post would be posted on the 16th
[00:40] Haplo But of course if you kept that growth up you would argue that the 2015 post would be posted on the 23rd.
[00:41] Mogs It wouldn't be the 23rd because it's reusing the previous one
[00:41] Haplo What do you mean, reusing the previous one?
[00:41] Mogs The cumulative growth is what would have to be found between the three years
[00:41] Haplo I
[00:42] Haplo Yes, that's what I've been saying :|
[00:42] Haplo You can't actually predict 2015's until you have more than two dates
[00:42] Mogs We're assuming the drift goes 2-5-7
[00:42] Mogs er
[00:42] Mogs no, now I'm confused
[00:42] Haplo What? :|
[00:43] Haplo The drift is 7 days
[00:43] Haplo Because 2012 was 2 days delay
[00:43] Haplo 2013 was 9 days
[00:43] Mogs which means it wouldn't be another 7 days
[00:43] Haplo Which is a delay growth of 7 days
[00:43] Haplo Then you want to do percentile growth
[00:44] Mogs We're going from Day Zero (first year) to the second (second year) to the ninth (third year)
[00:44] Haplo Basically, your delay growth rate for 2013 was 350%.
[00:44] Haplo By that calculation, your 2014 Stats post would be posted on the 24th.
[00:45] Mogs am I like trying to do exponential instead or something
[00:46] Mogs Basically how I'm seeing it
[00:48] Haplo What are you actually trying to do? :|
[00:48] Haplo You're trying to work out the delay on which you post the Stats post for 2014, right?
[00:52] Mogs yes
[00:55] Mogs 0<---------Day Zero
[00:55] Mogs |> 2<-----Delay 1
[00:55] Mogs ||>2<-----Increase of 2 (2-0)
[00:55] Mogs 2<---------Day posted
[00:55] Mogs |> 7<-----Delay 2
[00:55] Mogs ||5<-------Increase of 5 (7-2)
[00:55] Mogs 9<---------Day Posted
[00:55] Mogs |>7<------Projected Delay 3 (Because Cumulative increase of 2+5)
[00:55] Mogs ||>9<-----Increase of 9 (16-7)
[00:55] Mogs 16<--------Day Posted
[00:55] Mogs |>?<-------Delay desired
[00:59] Mogs extrapolation! *jazz hands*
[01:02] Mogs my brain is sad
[01:03] Haplo No no
[01:03] Haplo Doing that sort of cumulative growth rate is problematic
[01:04] Haplo Because you're not measuring delay growth
[01:04] Haplo You're measuring total delay
[01:04] Haplo And that's warping your results.
[01:05] Mogs but it is technically cumulative, right
[01:05] Haplo Yeah, but you can't use this to predict anything
[01:06] Mogs I'm still saying I was close
[01:06] Haplo 0->2->5->9
[01:06] Haplo Is a delay chart of 0->2->3->4
[01:07] Haplo So following -that- growth chart, you'd post your Stats post for 2014 on the 14th.
[01:08] Mogs Numbers broke me
[01:08] Mogs I am broken
[01:08] Haplo What you've done
[01:08] Mogs they are a harsh mistress
[01:09] Haplo Is the equivalent of saying that Person A was born in 2000, had an Age Growth of 1 in 2001, and 1 in 2002, and so in 2003 he'll grow by 2 years.
[01:09] Mogs but I did it in the most fucking complicated way possible
[01:09] Haplo Yup.
[01:09] Mogs goddamnit
[01:11] Mogs Do you mind if I post this as a chatlog in a post as sort of an "Outtake" post tonight since I'm obviously not going to get anymore math done
[01:11] Haplo | This.
[01:11] Haplo Sure, go ahead.
[01:11] Haplo But read that comic
[01:11] Mogs oh my god
[01:11] Mogs that is me
[01:12] Mogs I am that girl
[01:12] Mogs minus being a girl
[01:14] Haplo Just another fine chapter in the book of how I am better at everything
[01:14] Mogs I will cut you
[01:15] Haplo But who'll fix your maths then?

I would like to say that no matter how many times I threaten bodily harm to one of my best friends, I insist that I'm not a bad person.  Also, yes, I fiddled with the timestamp.

No comments:

Post a Comment