Support the Arctic Sea Ice Forum and Blog

Author Topic: September Predictions Challenge  (Read 6756 times)

Steven

  • ASIF Citizen
  • Posts: 462
    • View Profile
  • Liked: 58
  • Likes Given: 12
Re: September Predictions Challenge
« Reply #50 on: July 11, 2018, 08:21:42 PM »
June NSIDC cliff is absent or in the wrong direction but the data where it shoots up is compromised by satellite tests, so I'm just treating it as absent rather than an upwards cliff in June cancelled by a downwards one in July.

I don't see a reason to ignore the NSIDC data  (apart from the obvious outlier on 27 June).  In fact, JAXA is in good agreement with NSIDC: Both of them have very high compactness in the second half of June, and a strong decrease of compactness in the first week of July:



May data from Schroeder was 5.3 with confidence given by a 0.5 standard deviation.
June data is 4.7 +- 0.5 but I'm not sure if thats a standard deviation or a confidence range.

It's called forecast error in the Schroeder et al. 2014 paper.  It's obtained by simulating the "predictions" that the model would have generated in previous years, using only data that were available at the moment of the "prediction".  E.g. they simulate a prediction for September 2010 by using the regression equation over 1979-2009 and then feeding the May/June 2010 data into that equation.  They repeat that procedure for every year from 1984 to 2013 and then calculate the standard deviation of the errors of the simulated "predictions", giving them 0.44M km2 in the paper.  I guess the 0.5M km2 is obtained by including the last 4 years of data too (up to 2017) and perhaps rounding to the nearest multiple of 0.1M km2.

Brigantine

  • ASIF Lurker
  • Posts: 98
    • View Profile
  • Liked: 13
  • Likes Given: 53
Re: September Predictions Challenge
« Reply #51 on: July 12, 2018, 12:23:36 AM »
95% CI for prediction = 3.68 to 5.74
Thanks. Definitely sticking with medium confidence then!

uniquorn

  • ASIF Citizen
  • Posts: 480
    • View Profile
  • Liked: 165
  • Likes Given: 30
Re: September Predictions Challenge
« Reply #52 on: July 12, 2018, 11:52:08 PM »
uniquorn
JAXA min 3.5-4.0; medium
NSIDC av 3.75-4.25; medium

Steven

  • ASIF Citizen
  • Posts: 462
    • View Profile
  • Liked: 58
  • Likes Given: 12
Re: September Predictions Challenge
« Reply #53 on: July 14, 2018, 12:30:38 PM »
JAXA minimum extent:      4.25-4.75,  Medium
NSIDC September extent:  4.75-5.25,  Medium

This is based on the calculations discussed here.  Updating those calculations to use the latest July data, the calculated values are slightly lower than when using the June data, but still within the same bin.

Neven

  • Administrator
  • ASIF Royalty
  • *****
  • Posts: 5950
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 223
  • Likes Given: 155
Re: September Predictions Challenge
« Reply #54 on: July 14, 2018, 01:57:50 PM »
JAXA minimum extent:      4.25-4.75,  Medium
NSIDC September extent:  4.5-5.0,  Medium

My predictions for June didn't have any confidence levels, because I didn't get how it works. Can I still change them to 'low'?

Quote
Neven
June
NSIDC: Between 4.5 and 5.0 million km2, None
JAXA: 4.0 - 4.5 M km^2, None
Compare, compare, compare

Juan C. García

  • ASIF Middle Class
  • Posts: 843
    • View Profile
  • Liked: 132
  • Likes Given: 190
Re: September Predictions Challenge
« Reply #55 on: July 14, 2018, 02:14:55 PM »
JAXA minimum extent:      4.25-4.75,  Medium
NSIDC September extent:  4.5-5.0,  Medium

Same as Neven.
Just a coincidence. I did not copy him.  ;)
________________________________________
Edit:
This month, the level of confidence is medium.
________________________________________
Last month:
June:
JAXA - 3.75-4.25, Medium
NSDIC - 4.0-4.5, Medium


I voted the same way that jdallen.
Confidence? I think that there is a low confidence for everybody, including me. But let's say that I could expect one bin up, but several bins down.
« Last Edit: July 14, 2018, 06:50:12 PM by Juan C. García »
Which is the best answer to Sep-2012 ASI lost (compared to 1979-2000)?
50% [NSIDC Extent] or
73% [PIOMAS Volume]

Volume is harder to measure than extent, but 3-dimensional space is real, 2D's hide ~50% thickness gone.
-> IPCC/NSIDC trends [based on extent] underestimate the real speed of ASI lost.

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #56 on: July 14, 2018, 03:16:26 PM »
JAXA minimum extent:      4.25-4.75,  Medium
NSIDC September extent:  4.5-5.0,  Medium

My predictions for June didn't have any confidence levels, because I didn't get how it works. Can I still change them to 'low'?

Quote
Neven
June
NSIDC: Between 4.5 and 5.0 million km2, None
JAXA: 4.0 - 4.5 M km^2, None

I'm treating None as Very Low, (in all cases, not just yours), which is the option that I put in for those that didn't get it. Its far too late to change.

Almost too late for this month too, a little under 10 hours left from this post.

Neven

  • Administrator
  • ASIF Royalty
  • *****
  • Posts: 5950
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 223
  • Likes Given: 155
Re: September Predictions Challenge
« Reply #57 on: July 14, 2018, 04:07:52 PM »
Very low is perfect.  :)
Compare, compare, compare

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #58 on: July 23, 2018, 06:58:37 PM »
Compilation should now be up to date. Among those that have entered for both months, there is a general trend to higher predictions, almost everyone being 1 or 2 bins higher in July than June but the confidence is very little changed, almost all sticking at medium.

The July SIPN report shows a modest uptick, from 4.6 to 4.7, but almost all the movement is due to statistical methods where the change is pretty similar to the changes people made here.

In June I picked out: Rob Dekker's SIPN entry as  4.5 - 5.0  M  here, Nico Sun looks like 4.0 - 4.5 M but there are inconsistencies in the confidence measures he quotes, and it might be VL or H, Slater method 4.75 - 5.25 M, UCL model 4.0 - 4.5 L and the model of the Met Office 4.75 - 5.25 L.

For July I'd rate Rob Dekker at 5.0-5.5 H (borderline M), Nico Sun 4.25-4.75 M (same problem with interpreting the confidence measures as June), Slater method 4.75-5.25 M, UCL and Met Office did not made a new assessment for July. Random Weather I'd put as 4-4.5 H (also borderline M, there's an element of judgement in translating a model into an entry, so unless the poster makes that judgement in their post, I'm just treating it as commentary rather than an entry). Schroeder I'm putting at 5.0-5.5 M for June and 4.5-5.0 M for July, although that method ought to be have narrower confidence limits in July, and I went H in my own entry here on that basis, their SIPN entry did not change. The predictomat I rate at 4.5-5.0 M.

Looking forward to August, my preferred method switches to Slater which is projecting around 4.6 for Sept 11th at the moment. http://cires1.colorado.edu/~aslater/SEAICE/


gerontocrat

  • ASIF Governor
  • Posts: 3362
    • View Profile
  • Liked: 314
  • Likes Given: 5
Re: September Predictions Challenge
« Reply #59 on: July 23, 2018, 10:11:48 PM »
Looking forward to August, my preferred method switches to Slater which is projecting around 4.6 for Sept 11th at the moment. http://cires1.colorado.edu/~aslater/SEAICE/

The Slater figure is for the NSIDC DAILY extent, n'est-ce-pas? Which is usually a bit higher than the JAXA daily extent? My incredibly crude use of 10 previous years average remaining melt produces a JAXA daily extent minimum of 4.59 million km2. Perhaps I should apply the same method to NSIDC daily data. Yes I will.
"Para a Causa do Povo a Luta Continua!"
"And that's all I'm going to say about that". Forrest Gump
"Damn, I wanted to see what happened next" (Epitaph)

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #60 on: July 24, 2018, 08:02:54 PM »
Looking forward to August, my preferred method switches to Slater which is projecting around 4.6 for Sept 11th at the moment. http://cires1.colorado.edu/~aslater/SEAICE/

The Slater figure is for the NSIDC DAILY extent, n'est-ce-pas? Which is usually a bit higher than the JAXA daily extent? My incredibly crude use of 10 previous years average remaining melt produces a JAXA daily extent minimum of 4.59 million km2. Perhaps I should apply the same method to NSIDC daily data. Yes I will.

The application of it on the link I quoted is a daily forecast for 50 days in advance. It can also be used for forecasts over other periods, (and is done so for SIPN), by those with access to the model details. For the August poll here, I look at what it has said 50 days in advance across September, and subject to a couple of reality checks, roughly average it.

gerontocrat

  • ASIF Governor
  • Posts: 3362
    • View Profile
  • Liked: 314
  • Likes Given: 5
Re: September Predictions Challenge
« Reply #61 on: July 24, 2018, 08:58:37 PM »
Looking forward to August, my preferred method switches to Slater which is projecting around 4.6 for Sept 11th at the moment. http://cires1.colorado.edu/~aslater/SEAICE/

The Slater figure is for the NSIDC DAILY extent, n'est-ce-pas? Which is usually a bit higher than the JAXA daily extent? My incredibly crude use of 10 previous years average remaining melt produces a JAXA daily extent minimum of 4.59 million km2. Perhaps I should apply the same method to NSIDC daily data. Yes I will.

The application of it on the link I quoted is a daily forecast for 50 days in advance. It can also be used for forecasts over other periods, (and is done so for SIPN), by those with access to the model details. For the August poll here, I look at what it has said 50 days in advance across September, and subject to a couple of reality checks, roughly average it.
The average date for the JAXA minimum (last 10 years) is Sep 12th - 50 days from now. The Slater 50 day projection is likely to be very close to the minimum this year (unless the oddness of this year produces a very mild late summer and fall?) ?
"Para a Causa do Povo a Luta Continua!"
"And that's all I'm going to say about that". Forrest Gump
"Damn, I wanted to see what happened next" (Epitaph)

Phil42

  • ASIF Lurker
  • Posts: 24
    • View Profile
  • Liked: 25
  • Likes Given: 101
Re: September Predictions Challenge
« Reply #62 on: July 30, 2018, 01:51:08 PM »
I'm taking a little bit of a different approach here to make my (August?) prediction. I looked at the JAXA minimum extent ranks the previous years placed in (see graph below) and from there took the average rank, which is 4.21 from 1990-2017.

For 2018 to be placed at exactly this average rank of 4.21, the JAXA minimum extent would have to be 4'344'239km2.

So my prediction is
4.0 - 4.5 JAXA
4.5 - 5.0 NSIDC

(I am aware that is a really bad method to make a prediction but I was just wondering what the outcome of the calculation would be and though I might as well share it  ;))

Paddy

  • ASIF Citizen
  • Posts: 430
    • View Profile
  • Liked: 4
  • Likes Given: 3
Re: September Predictions Challenge
« Reply #63 on: July 31, 2018, 11:01:26 AM »
I think for the August round we should generally be able to state things with a higher level of confidence, as we're now only about 45 days from the minimum. And although the range of minima for the past ten years has been very large, it really doesn't look like this year will be on an extreme end of that range.

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #64 on: July 31, 2018, 05:53:55 PM »
I think for the August round we should generally be able to state things with a higher level of confidence, as we're now only about 45 days from the minimum. And although the range of minima for the past ten years has been very large, it really doesn't look like this year will be on an extreme end of that range.

August Polls have been up long enough to get a fair number of entries so even though they aren't stickied yet, I'm opening this Challenge for entries based on them. August 12th (BST) is the final day to post here.

I expect there will be some upwards movement in confidence, though its quite a stretch to go from H to VH for me, that overconfidence penalty is really quite severe. I'm still happy with my picks from last month's polls though, so I'll be thinking about whether I can stretch on the NSIDC average as the deadline approaches.

jdallen

  • ASIF Governor
  • Posts: 2757
    • View Profile
  • Liked: 62
  • Likes Given: 99
Re: September Predictions Challenge
« Reply #65 on: July 31, 2018, 06:02:04 PM »
If a anything, the last few days have driven home to me just how volatile and unpredictable conditions are. Sticking with medium.
This space for Rent.

Paddy

  • ASIF Citizen
  • Posts: 430
    • View Profile
  • Liked: 4
  • Likes Given: 3
Re: September Predictions Challenge
« Reply #66 on: July 31, 2018, 06:58:14 PM »

I expect there will be some upwards movement in confidence, though its quite a stretch to go from H to VH for me, that overconfidence penalty is really quite severe. I'm still happy with my picks from last month's polls though, so I'll be thinking about whether I can stretch on the NSIDC average as the deadline approaches.

Fair point if you're already going H.  But personally, I think it's about time I went up from my L.

If a anything, the last few days have driven home to me just how volatile and unpredictable conditions are. Sticking with medium.

Fair point as well; but there's now less time left for that volatility to happen.

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #67 on: August 01, 2018, 12:57:07 PM »
If a anything, the last few days have driven home to me just how volatile and unpredictable conditions are. Sticking with medium.

This is why I like the Slater model. It gives me a method of assessing the implications of concentration changes now, for the extent in 50 days time. And at the moment its completely serene about them, ticking the amount of ice it expects to survive up slightly over the past week. i.e. ice that was going anyway is going anyway, but the places that are likely to be marginal in September are looking a teeny bit more solid. However, maybe next week those marginal places start getting eaten ... which is why I'm both happy with my July pick, and running down the clock before making the August one.

EgalSust

  • ASIF Lurker
  • Posts: 36
    • View Profile
  • Liked: 1
  • Likes Given: 1
Re: September Predictions Challenge
« Reply #68 on: August 01, 2018, 03:03:45 PM »
August
JAXA - 4.25-4.75, High
NSIDC - 4.75-5.25, High

for now... sticking with the following Slater method, for now at least

Slater probabilistic is showing very weird behaviour: a late August minimum and a sharp uptick after mid-September:


Tor Bejnar

  • ASIF Upper Class
  • Posts: 2107
    • View Profile
  • Liked: 82
  • Likes Given: 19
Re: September Predictions Challenge
« Reply #69 on: August 01, 2018, 03:47:17 PM »
I "predict" (defined as "guess") ASI extent, area and volume will just beat out the current records in September.  ...
I now guess the September minimums will be just above the current records, so (to enumerate the guesses) I choose the lowest bin that starts above the record low for each measure (extent, area, volume).  I still say guesses should be recognized as having no confidence.  As of yet unknowable weather, of course, will be a significant contributor to making or breaking every guess, prediction or sure thing declaration.
Arctic ice is healthy for children and other living things.

uniquorn

  • ASIF Citizen
  • Posts: 480
    • View Profile
  • Liked: 165
  • Likes Given: 30
Re: September Predictions Challenge
« Reply #70 on: August 02, 2018, 02:39:39 PM »
August
JAXA - 3.5-4.0, medium
NSIDC - 3.75-4.25 medium
Warm winds and waves

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #71 on: August 09, 2018, 04:02:56 PM »
The clock is ticking and its time I put my entry in. While 2 weeks ago I was happy with the bin I picked last month, Slater has shifted fairly dramatically compared to anomaly persistence during that time, my reading of Wipneus' home brew areas backs up that shift, and I'm going up a bin.

VH or H that is the question. Breakeven expectation if only being 1 bin out is considered happens at a 75% chance of hitting the right bin. (i.e. at 80% correct, 10% 1 high, 10% 1 low, VH comes out ahead) While I am that confident on the high side, I'm not on the low side, so I'll have to stick at H.

NSIDC: 4.75-5.25, High
JAXA: 4.25-4.75, High

Because I'm not that confident on the low side, I've put JAXA two bins lower, rather than my normal 1 bin lower than NSIDC. (historical average is 0.3 lower, slightly more than 1 bin)


Brigantine

  • ASIF Lurker
  • Posts: 98
    • View Profile
  • Liked: 13
  • Likes Given: 53
Re: September Predictions Challenge
« Reply #72 on: August 10, 2018, 01:52:22 AM »
JAXA - 4.00-4.50, Medium
NSIDC - 4.25-4.75, Medium
July entry:
JAXA - 4.25-4.75 Medium
NSIDC - 4.50-5.00 Medium
August entry: Same bin as June, but with higher confidence
JAXA - 4.00-4.50, High
NSIDC - 4.50-5.00, High


And where's that predict-o-matic up to now, including confidence intervals?
Found it! Mean is 4.23, 95% range is 3.70-4.77. Thanks Ned!
My estimate is slightly higher than that, but not by enough to pick a higher bin. Reason being highish concentration & potential for dispersion.

[Update - joined the bandwagon and raised the NSIDC entry by one bin so that it's 2 bins above JAXA]
« Last Edit: August 12, 2018, 02:11:49 AM by Brigantine »

Paddy

  • ASIF Citizen
  • Posts: 430
    • View Profile
  • Liked: 4
  • Likes Given: 3
Re: September Predictions Challenge
« Reply #73 on: August 10, 2018, 07:19:04 AM »
JAXA: 4.0 to 4.5, Medium
NSIDC: 4.5 to 5.0, Medium

Was very tempted to go High this time...

EDIT: my previous votes for comparison:

Quote
June
JAXA - 3.75-4.25, Low
NSDIC - 4.25-4.75, Low
July
JAXA - 4.0-4.5, Low
NSDIC - 4.5-5.0, Low
« Last Edit: August 12, 2018, 07:57:30 AM by Paddy »

Viggy

  • ASIF Lurker
  • Posts: 17
    • View Profile
  • Liked: 11
  • Likes Given: 7
Re: September Predictions Challenge
« Reply #74 on: August 10, 2018, 12:09:33 PM »
JAXA: 3.25 to 3.75, High
NSIDC: 3.75 to 4.25, Very High

I don't know if my rationale is valid or not but accumulated cyclonic energy this hurricane season has been running quite high and all that heat and moisture eventually ends up polewards.

http://tropical.atmos.colostate.edu/Realtime/

I'd imagine that transfer of energy can only exacerbate a GAC like event late in the melt season. Expecting September to be very slow in its re-freezing because of all this energy.
« Last Edit: August 10, 2018, 09:38:20 PM by Viggy »

Tetra

  • ASIF Lurker
  • Posts: 31
    • View Profile
  • Liked: 7
  • Likes Given: 1
Re: September Predictions Challenge
« Reply #75 on: August 10, 2018, 10:19:15 PM »
JAXA: 4.0 to 4.5, Medium
NSIDC: 4.5 to 5.0, Medium

Was very tempted to go High this time...

 8)

I’ll go for those same odds.

Steven

  • ASIF Citizen
  • Posts: 462
    • View Profile
  • Liked: 58
  • Likes Given: 12
Re: September Predictions Challenge
« Reply #76 on: August 11, 2018, 12:45:35 PM »
JAXA minimum extent:      4.0-4.5, High
NSIDC September extent:  4.5-5.0,  High

This is slightly lower than my vote in the July polls, but higher than in the June polls.

NSIDC extent is currently 6th lowest or so.  Extent trajectories for previous years suggest about 4.6 million km2 for the NSIDC September extent.  On the other hand, Slater's method seems to suggest about 4.9 million km2 for the NSIDC September extent.  In both cases this is in the 4.5-5.0 bin.

For JAXA extent I choose the 4.0-4.5 bin, which is two bins below NSIDC.  The difference between JAXA minimum and NSIDC September extent was about 0.3 million km2 on average in the past 15 years.  But in the past 6 years (since JAXA started to use AMSR2 data) it was always more than 0.3, and almost 0.4 million km2 on average.

Neven

  • Administrator
  • ASIF Royalty
  • *****
  • Posts: 5950
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 223
  • Likes Given: 155
Re: September Predictions Challenge
« Reply #77 on: August 11, 2018, 01:15:34 PM »
JAXA minimum extent:      4.0-4.5,  Medium
NSIDC September extent:  4.5-5.0,  High

In June it was:
JAXA: 4.0-4.5, None
NSIDC: 4.5-5.0, None

And for July it was:
JAXA: 4.25-4.75, Medium
NSIDC: 4.5-5.0, Medium

My NSIDC prediction has been very consistent!  ;)
Compare, compare, compare

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #78 on: August 12, 2018, 01:30:45 PM »
Last day for challenge entries. Deadline is about 11 hours from this post.

liefde

  • ASIF Lurker
  • Posts: 41
  • ice melt will trump donald
    • View Profile
  • Liked: 4
  • Likes Given: 1
Re: September Predictions Challenge
« Reply #79 on: August 12, 2018, 11:39:08 PM »
JAXA ASI September daily minimum: Between 3.5 and 4.0 million km^2, Very High
NSIDC SIE September average: Between 4.25 and 4.75 million km^2, Very High

Purely based on a climatic heli-view, based on good friend Jörg (Kachelmann)'s model input.
And with Aug. 16th ECMWF HD predictions taking another pull at the NH, with Greenland and Svalbard meltwater getting warmer than ever, and SST anomalies of all oceanic surroundings at record high for August, thick new heat blankets of CH4 at surface level all over EurAsia. The energy has to go somewhere.

Wherestheice

  • ASIF Citizen
  • Posts: 162
    • View Profile
  • Liked: 32
  • Likes Given: 10
Re: September Predictions Challenge
« Reply #80 on: August 16, 2018, 01:48:15 AM »
Hopefully the jaxa data can come in here soon, so we can make some good guesses on what the minimum will be.
"When the ice goes..... F***

Juan C. García

  • ASIF Middle Class
  • Posts: 843
    • View Profile
  • Liked: 132
  • Likes Given: 190
Re: September Predictions Challenge
« Reply #81 on: August 18, 2018, 05:15:27 PM »
Last day for challenge entries. Deadline is about 11 hours from this post.

Seems that I like to break the rules...  ;D  :-[
Sorry, I forgot to put here my numbers:

June:
   JAXA: 3.75 - 4.25, Low
   NSIDC: 4.0 – 4.5, Low
July:
   JAXA: 4.25 – 4.75 Medium
   NSIDC: 4.5 – 5.0 Medium
August:
   JAXA: 4.0 – 4.5 High
   NSIDC: 4.25 – 4.75 Medium

I can see that NSIDC September average goes a little above 4.75M km2. It will depend of the freezing on Sep. 15th-30th.
Which is the best answer to Sep-2012 ASI lost (compared to 1979-2000)?
50% [NSIDC Extent] or
73% [PIOMAS Volume]

Volume is harder to measure than extent, but 3-dimensional space is real, 2D's hide ~50% thickness gone.
-> IPCC/NSIDC trends [based on extent] underestimate the real speed of ASI lost.

Brigantine

  • ASIF Lurker
  • Posts: 98
    • View Profile
  • Liked: 13
  • Likes Given: 53
Re: September Predictions Challenge
« Reply #82 on: September 21, 2018, 10:24:47 PM »
It's looking very likely JAXA will be in the 4.25-4.50 range, and NSIDC will be either side of 4.75.

The rest of us are still waiting for NSIDC to make its mind up, but Neven can be scored now.

He got 6 out of 6! Jackpot. +20 points  ;D
(would have been an even higher score if he picked a confidence level in June)


I picked all the same bins as him except for June NSIDC (I picked 4.25-4.75). So I got at least 5/6 right. And I had higher confidence levels in June & for August JAXA.

If NSIDC comes in over 4.75, I get +26 points, if it comes in under, I get +28. It's more luck than skill, I swear!  :o Maybe I should quit while I'm ahead and not play in the freezing season...  :P

Paddy

  • ASIF Citizen
  • Posts: 430
    • View Profile
  • Liked: 4
  • Likes Given: 3
Re: September Predictions Challenge
« Reply #83 on: September 21, 2018, 11:22:55 PM »
I got 7 points for Jaxa (1 point for missing by 1 bin in June on Low, 2 points for a hit in July on Low and 4 points for a hit in August on Medium). As expected, going for Low confidence dragged down the scores, but it felt like the most honest description, especially in June. Probably 7 or 8 for NSIDC, making it probably 14 or 15 points.

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #84 on: September 25, 2018, 12:41:04 PM »
August predictions editted into the top post, if I missed any, let me know.

One titbit from August SIPN: Slater model went for 4.75 which might well be taken as 4.5-5.0 VH, though there's not really enough information given on the uncertainty to be sure it should be VH rather than H.

I'll tot up the scores once the NSIDC September report is out.

Richard Rathbone

  • ASIF Citizen
  • Posts: 455
    • View Profile
  • Liked: 10
  • Likes Given: 6
Re: September Predictions Challenge
« Reply #85 on: October 06, 2018, 01:43:50 PM »
JAXA daily minimum was 4.46
NSIDC Sept average was 4.71

Scores (details in 1st post if you want to check my scoring)

Richard Rathbone 24
Paddy 15
Brigantine 28
Neven 20
EgalSust 24
jdallen 4
Daniel B. 10
Stephan 10
Viggy -32
gerontocrat 0
RealityCheck 8
Juan C. Garcia 12
uniquorn 6
Steven 22
Tetra 8
liefde 0

Most regular competitors either hit the bull's eye or missed by at most one bin on all entries.
Almost every use of VH was overconfident and took the -10 penalty, although plenty of entrants would have got lucky if they had used it, but it would have turned out badly for me if I had gone VH when I was thinking about it.

Brigantine

  • ASIF Lurker
  • Posts: 98
    • View Profile
  • Liked: 13
  • Likes Given: 53
Re: September Predictions Challenge
« Reply #86 on: October 07, 2018, 12:38:53 PM »
Interesting that NSIDC was exactly one bin above JAXA - 0.25. So both were just 0.04 below the bin boundary.

A look at the confidence levels used by the 5 most successful participants:

28 Brigantine - 4x Medium, 2x High, maximum luck
24 Richard Rathbone - 4x High, 2x Medium, estimates drifted up 2-3 bins over time
24 EgalSust - 4x High, 2x Medium, narrowly missed out on an incredible score of 32 (if NSIDC were 0.04 higher)
22 Steven - 4x Medium, 2x High, estimates followed the predict-o-matic - up 2-3 bins in July then 1 bin back down
20 Neven - average confidence of Low-Medium (partly by accident), maximum bin accuracy.

Daniel B. and Stephan both did very well in June (10 points, called NSIDC dead accurately with High Confidence), but then didn't come back to this thread in July or August.
« Last Edit: October 07, 2018, 01:08:51 PM by Brigantine »

Stephan

  • ASIF Citizen
  • Posts: 182
    • View Profile
  • Liked: 27
  • Likes Given: 13
Re: September Predictions Challenge
« Reply #87 on: October 07, 2018, 08:28:59 PM »
I would have stayed with my bets throughout the summer and therefore I didn't repeat them. Should I have done so?

Brigantine

  • ASIF Lurker
  • Posts: 98
    • View Profile
  • Liked: 13
  • Likes Given: 53
Re: September Predictions Challenge
« Reply #88 on: October 07, 2018, 08:52:10 PM »
I would have stayed with my bets throughout the summer and therefore I didn't repeat them. Should I have done so?
If you had made the same entry in July and in August, you would have won!

Stephan

  • ASIF Citizen
  • Posts: 182
    • View Profile
  • Liked: 27
  • Likes Given: 13
Re: September Predictions Challenge
« Reply #89 on: October 09, 2018, 06:11:37 PM »
OK. As a newbie here in this forum and as a non climate scientist I feel honoured.
If there will be a comparable thread in 2019 I will not forget to repeat my June predictions in July and August!