• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Official May 2007 NPD Prediction Thread

D

DSWii60

Member
Mine:
[PS2] 155k
[PSP] 180k
[PS3] 85k
[360] 175k
[GBA] 105k
[NGC] 9k
[NDS] 435k
[WII] 370k

Actual:
[NDS] 423,150
[WII] 338,278
[PSP] 221,120
[PS2] 187,765
[360] 154,932
[PS3] 81,604
[GBA] 80,554
[NGC] 10,728

Ah not a good month for me, 20-30k off on everything apart from PS3 and NGC.
 
V

Vagabundo

Member
Not bad I got #39...

I should fess up as I did not even think about the figures much. Just looks at a few of the consistent high rankers and munched the results up or down a couple of thou...

I know, I am a leech.
 
T

the thoroughbred

Member
Damn, I was sure I'd have done better this time.

69. the thoroughbred - 159,747. Meh not bad though. I'm happy with my position, so is she (neogaf).:lol :lol
 
D

donny2112

Member
apujanata said:
Can you expand the list so that it cover Top 30,
Click to expand...

Code: 
    May-07                    Last 3 months                2007
1.  The Sphinx       162.48   The Sphinx          392.65   BuzzJive            616.13
2.  Dice Man         146.12   BuzzJive            389.05   sonycowboy          616.05
3.  Fuzzy            142.24   apujanata           379.49   argon               608.37
4.  cvxfreak         141.94   Heidir              371.27   GhaleonEB           601.32
5.  LiquidMamba      141.94   Jokeropia           366.74   donny2112           600.42
6.  TwinsDad         141.18   Tf53                365.33   Clever Pun          589.9
7.  davepoobond      139.43   Clever Pun          365.15   JoshuaJSlone        587.83
8.  kassatsu         137.92   Branduil            362.84   rage1973            576.92
9.  rage1973         135.19   Bildi               360      duderon             576
10. borghe           134.75   TheKurgan           359.64   starship            573.4
11. Chiggs           133.87   Gekkonidae          358.63   Orgen               569.86
12. duk              132.13   argon               358.37   Heidir              566.63
13. JDSN             131.85   Haunted One         357.69   Kurosaki Ichigo     565.29
14. Nicktals         131.26   DSWii60             357.15   bycha               557.56
15. gollumsluvslave  130.47   Parl                357.08   Lock                555.09
16. fanduck          130.24   F#A#Oo              356.93   hadareud            539
17. jey_16           129.5    donny2112           355.99   beermonkey@tehbias  524.63
18. bud              129.14   duderon             355.11   Earthstrike         515.21
19. titiklabingapat  129.09   sonycowboy          354.91   PhoenixDark         497.36
20. bycha            129.05   Odysseus            353.73   The Sphinx          487.01
21. Branduil         128.96   Pope Benedict XVI   353.6    Jokeropia           480.15
22. sonycowboy       128.75   starship            353.37   DayShallCome        475.56
23. Saitou           128.68   DayShallCome        353.32   Chibbi Brady        463.73
24. MikeMyers        128.6    Mariah Carey        352.5    JJConrad            459.74
25. Chibbi Brady     128.52   DeaconKnowledge     352.04   TheKurgan           458.84
26. John Dunbar      128.5    felipeko            350.61   Branduil            456.27
27. NinSoX           128.43   GhaleonEB           349.91   Frankfurter         453.29
28. verplant         128.41   rage1973            349.59   Ganondorf>Link      443.23
29. TheKurgan        128.22   beermonkey@tehbias  349.26   fanduck             442.53
30. F#A#Oo           128.19   Orgen               347.69   AniHawk             429.55

cvxfreak said:
I didn't know scoring had been in progress earlier. :(
Click to expand...

It wasn't, but as I had the results from January on, I just backfit it.

test_account said:
Those scores in post #661, what do the number indicates?
Click to expand...

The sum of the absolute differences between predicted and actual.
 
D

DayShallCome

Member
Thanks for all the hard work guys (Donny, Cheese, SonyCowboy). This ranking system is pretty solid - though I agree that the 20 point bonus is far too generous. Perhaps 5 or 10 points instead?

Edit: The top 30 listings are great, as it gives people who are not at the top incentive to keep at it.
 
F

Frankfurter

Member
Perhaps I'm missing sth., but how is it possible that there are only like ~10 people better than the GAF average, but ~200 worse (in the posts 661 and 664). Shouldn't the average be somewhere more in the middle?

Btw. yeah, I'm #27 on the 2007 list :D

Btw. btw.: thanks for all the work you guys put into the formulars, the predictions threads, the results etc., it's really, really appreciated :)
 
E

Evlar

Banned
Frankfurter said:
Perhaps I'm missing sth., but how is it possible that there are only like ~10 people better than the GAF average, but ~200 worse (in the posts 661 and 664). Shouldn't the average be somewhere more in the middle?

Btw. yeah, I'm #27 on the 2007 list :D

Btw. btw.: thanks for all the work you guys put into the formulars, the predictions threads, the results etc., it's really, really appreciated :)
Click to expand...
That happens because some people over-estimate the sales for a particular piece of hardware (say, the 360). Other people will under-estimate the 360. However, when you average them, the over- and under-estimates balance each other out and you get a consensus value that's closer than many of the individual estimates.

It proves that while the individuals on GAF can have biases and distorted expectations for sales, GAF as a whole, as a collective, has a very good bead on the hardware market.
 
D

DSWii60

Member
I'm 14th in the last 3 months :) (the only 3 months I've predicted)

If only I had done better this month...
80th, ouch
 
B

BuzzJive

Member
Cheesemeister said:
Sure. Results with 360, PS3, and WII only...
Click to expand...

Once again - you rock.

donny2112 said:
2007
BuzzJive 616.13
Click to expand...

As do you...



I'm hanging on to my 2007 lead, but there's lots of folks right behind me. May was a bit brutal, but I'm crunching some numbers as we speak and preparing for a strong June.
 
S

sonycowboy

Member
BuzzJive said:
I'm hanging on to my 2007 lead, but there's lots of folks right behind me. May was a bit brutal, but I'm crunching some numbers as we speak and preparing for a strong June.
Click to expand...

Enjoy it while it lasts. It's almost impossible for me to be any closer than I am now.

2007
BuzzJive 616.13
sonycowboy 616.05

Next month, you're going down. I'm hot on your heels and with a special insider arrangement I'm going to activate, I'll be invincible! :D

-jk ;)
 
S

Sule

Member
Frankfurter said:
Perhaps I'm missing sth., but how is it possible that there are only like ~10 people better than the GAF average, but ~200 worse (in the posts 661 and 664). Shouldn't the average be somewhere more in the middle?

Btw. yeah, I'm #27 on the 2007 list :D

Btw. btw.: thanks for all the work you guys put into the formulars, the predictions threads, the results etc., it's really, really appreciated :)
Click to expand...

5zr15lj.jpg


A little explanation to go with the above.

Let's say NPD was one number, 500. Let's say the GAFerage was 450. Anybody that had a prediction between 450,000...01 and 549.99999999999..., has a smaller difference between their prediction and NPD than the GAFerage and NPD. In the rankings they would be above the GAFerage, because they have a smaller absolute difference. That's why it's possible for the GAFerage to be higher than the middle of the rankings. I think this could also be used for NPD, where you have multiple numbers, but obviously do the same thing x times.

At least this is what I could come up with, as an explanation. I could also do some math and explain in that way, but I'm playing Majora's Mask right now :)
 
G

GhaleonEB

Member
Yeah I still can't figure out what that's trying to say. :lol

79. GhaleonEB - 166,771


You know GAF is getting good at this when being off by that much lands me at #79. Wow.
 
D

donny2112

Member
GhaleonEB said:
You know GAF is getting good at this when being off by that much lands me at #79. Wow.
Click to expand...

I was less than 20K closer than you, but I'm 21 spots higher. Yea, it's pretty crazy now. :lol


I'm reading various prediction thread suggestions, but is there actual desire to change the system some?

1) End 1 day before NPD release.

2) Reduce points for correct order of systems from 15 to 10.

3) "Secret" ballots/website entry for predictions.

Also, I really liked Cheesemeister's summaries of the ranges/averages as the predictions were going on, but with how much better the GAF average did than most everybody else this month, I think that should probably not be posted until the predictions close. :)
 
N

Nicktals

Banned
Number 14 for my first month! Pretty good considering I didn't put any real thought into it! I just wish I had participated from the beginning of 2007.

EDIT:

NDS: +31,850
WII: +36,722
PSP: -29,120
PS2: -2,765
360: -2,932
PS3: -6,604
GBA: -554
NGC: -728

Did great on 360 ps2 and GBA. Nothing was too terrible.
 
F

fanduck

Member
Wow, I impressed myself this month :lol

sonycowboy said:
NOTE: Fanduck was the first to have all ordinal positions correct, so he would have won. Nobody else in the top 20 got all correct. The PSP > PS2 got almost everyone and for those it didn't, the PS3 > GBA got them.
Click to expand...

Well, technically, I guessed PSP = PS2, so that's kind of cheating, but I'll take the kudos anyway :D
 
C

Chibbi Brady

Member
Big thanks to everyone involved for all their hard work, really great to see everything working out so well. Really shocked to see the PSP sink nearly everyone, I was definitely expecting an increase in April, but when that didn't really materialize, I thought the price drop had little or no effect. Certainly wasn't expecting it to show up in May, really weird.

Oh and another few months of this kind of data and I think about 190 or so of us can apply for Pachter's job. :)
 
J

jamesinclair

Banned
Bah, I didnt know wed gotten so serious again. This month I just amde up numbers, instead of using a series of excel formulas. which is what kept me in the top 10 in the past.

That, and sampling the Nintendo hut
 
J

Jokeropia

Member
I guess I should be happy with the fifth place in the three month ranking, but I've still yet to match the performance I managed in my first prediction thread. :p (Where I finished third.)
 
S

starship

psycho_snake's and The Black Brad Pitt's B*TCH
donny2112 said:
1) End 1 day before NPD release.

2) Reduce points for correct order of systems from 15 to 10.

3) "Secret" ballots/website entry for predictions.

Also, I really liked Cheesemeister's summaries of the ranges/averages as the predictions were going on, but with how much better the GAF average did than most everybody else this month, I think that should probably not be posted until the predictions close. :)
Click to expand...
I agree with these suggestions. Actually I think 2 days before NPD release is better because usually Canadian numbers appear one day prior to US NPD release.
 
A

andthebeatgoeson

Junior Member
Cheesemeister said:
Rankings with precise numbers for everything but Xbox.
Click to expand...

Not that I'm a math wiz. But how is the average so low when the majority (193 predictions) are above the average. 10 below and 193 above the average.

Edited out. I'm slow.
Edit: Beaten. No matter what you think, GAF has thought it, debated it, made a joke about it, then jerked it to it.

The Sphinx said:
That happens because some people over-estimate the sales for a particular piece of hardware (say, the 360). Other people will under-estimate the 360. However, when you average them, the over- and under-estimates balance each other out and you get a consensus value that's closer than many of the individual estimates.

It proves that while the individuals on GAF can have biases and distorted expectations for sales, GAF as a whole, as a collective, has a very good bead on the hardware market.
Click to expand...

Well, should it be the absolute value of estimation? Maybe that's taken into account. Does this mean you could overestimate the 360 by 100K and underestimate the PS3 by 100K and still be considered accurate?
 
C

Cheesemeister

Member
skinnyrattler said:
Not that I'm a math wiz. But how is the average so low when the majority (193 predictions) are above the average. 10 below and 193 above the average.

Put another way:
If you have 100 data points and 99 of them are the number '5,000' and the remaining one was the number '5', you would expect the average to be closer to '5000' than 5. Am I missing something?
Click to expand...

See post 715. Perhaps "Aggregate" would be a better word to use?
 
S

sonycowboy

Member
Cheesemeister said:
See post 715. Perhaps "Aggregate" would be a better word to use?
Click to expand...

I think the problem for some folks is that there are multiple variables.

Example:


Prediction 1: 50 1500
Prediction 2: 150 500
Prediction 3: 125 1250
Prediction 4: 75 750
Actual: 100 1000
Average: 100 1000

In this case the average of each individual turned out to be the actual, so the "average" across multiple choices would turn out to be in first place even though there were 4 predictions that were in second through fifth place.

It's a perfect example of wisdom of the crowd. Individually, we're not that smart, but the concensus of each choice weeds out false hope, hype, and comes out to be pretty good. IMO, the larger the population, the better the GAF average will tend to be. Even more so as more people take it seriously. Even the ridiculous Gahiggidy and spwolf predictions can't have an effect because they're error gets spread out over 200 other predictions.
 
C

Cheesemeister

Member
skinnyrattler said:
Well, should it be the absolute value of estimation? Maybe that's taken into account. Does this mean you could overestimate the 360 by 100K and underestimate the PS3 by 100K and still be considered accurate?
Click to expand...

That's a much less accurate prediction evaluation system. By that measure, I could predict that the Xbox is going to sell 500k units in June and the DS will sell next to nothing, and still be called accurate because the sum of my hardware predictions was close to the actual hardware sum.
 
A

apujanata

Member
Cheese, I suggest that you only post the GAF average AFTER the prediction submission has been closed, and you have run the parser for the last time (to get the final position).

Your average calculation might made people revise their prediction.
 
F

Frankfurter

Member
Thanks for the explanation :) I thought that the "average" simply meant the average of all predictions (not all predictions on the different systems).

Btw. the descriptive GIF was a bit ... confusing :D I didn't see The Sphinx' post until after I saw that picture and read the post :lol
 
You must log in or register to reply here.
Top Bottom