Jump to content
Thunder?
Local
Radar
Pollen

April 2021 C.E.T. and EWP forecast contests


Recommended Posts

Posted
  • Location: Broxbourne, Herts
  • Weather Preferences: Snow snow and snow
  • Location: Broxbourne, Herts

    A little table of values of monthly mean CETs of May to December from years that are not far removed in aggregate terms for January to April  from where we are at the moment, followed by picks for the rest of this year that could end up in an annual CET of under 9C

    For a couple of years I've shaved a few points off to allow for the fact that there is not otherwise enough leverage to bring the overall CET below 9c

    I've shown where the monthly mean is either part of its coldest 10% or its warmest 10%.  I've then tried to look at the values in 2021 for the rest of the year that could bring us an annual CET under 9C .

    It's unlikely we will get another month that joins April as a coldest 10% member bearing in mind we've only had three such months all century and the last year we had more than once such month was in 1991 when both February and June registered in their coldest 10%.  So while the picks used for the rest of the year are never really particularly warm, they are avoiding the extremely cold options....

     

    image.thumb.png.153b14c7ed1ca5e697a8479bdb1c0a4f.png

    image.png.68ec42bf963182dcac677134ed00db17.png

     

     

    • Like 3
    Link to comment
    Share on other sites

    Spotted a post you think may be an issue? Please help the team by reporting it.
    Posted
  • Location: Rossland BC Canada
  • Location: Rossland BC Canada

    Looking at the 06z GFS, what's to stop May 2021 from being just as cold in relative terms as April or even some sort of Nov-Dec 2010 redux? 

    It looks really cold for two more weeks now. Could be as low as 7.5 C running CET values similar to 1996. A flip to quite warm seems fairly likely but will it be strong enough to keep May from ending up below 10C? 

    • Like 4
    Link to comment
    Share on other sites

    Posted
  • Location: Irlam
  • Location: Irlam

    April 2021 had a CET minimum of 1.0C but 3rd April-2nd May 2021 CET minimum has a provisional value of 0.7C

    The minimum value for the last 30 days has become even more remarkable over the last two mornings

    • Like 4
    Link to comment
    Share on other sites

    Posted
  • Location: Ludgershall, Wiltshire
  • Location: Ludgershall, Wiltshire
    55 minutes ago, Roger J Smith said:

    Looking at the 06z GFS, what's to stop May 2021 from being just as cold in relative terms as April or even some sort of Nov-Dec 2010 redux? 

     

    That would be a waste at this time of year!

    Link to comment
    Share on other sites

    Posted
  • Location: Powys Mid Wales borders.
  • Location: Powys Mid Wales borders.
    Posted (edited)

    It was quite a way off the driest april here that goes to 2011 then 2007.

    This month comes 7th at least.

    April is normally the driest month of the year anyway.

    Edited by Snowyowl9
    Link to comment
    Share on other sites

    Posted
  • Location: Rossland BC Canada
  • Location: Rossland BC Canada
    Posted (edited)

    Combined standings in CET and EWP contests

    Scoring in the EWP contest has been brought more in line with the CET by adjustments for accuracy and correct anomaly sign. But scores in both contest are not 1:1 measurements of skill, because in both contests you lose some points for having a later duplicate entry. Also there are late penalties although those might not be entirely inconsistent with skill level since later entries should have some advantage (probably less advantage than the points lost however). Therefore I have added a new feature this month, columns showing your rank for average error (C deg) and (mm) -- these do not take into account order of entry or late penalties and tell you how your are performing (as opposed to how you are playing the contest which is measured by rank in the contest under its scoring rules). If you see that your performance is better than your contest scoring, it means you are often later entering and that you are choosing duplicate entries. There is no clear math on whether it's good or bad to avoid duplicates. If you go 0.1 away from a popular choice, you only gain anything if the result is on that side of the popular choice -- if it's on the other side, you lose (more in the EWP because scoring levels only drop 2/3 of an interval for duplicates, in the CET it makes less difference to go 0.1 away although you have to factor in the intangible of how many entries will tie your entry on the other side of the result (e.g., you see 7.4 is popular so you go to 7.3, the outcome is 7.2 so you pass all the people at 7.4, but are still tied with the people at 7.1, order of entry applying in that pool). Since every time you choose to go off a popular entry the outcome is theoretically random in terms of gain or loss, it probably makes no difference in the long run. The table continues to blend CET-only entrants with those who play both contests. CET-only entrants have a scoring line in italics and their names are indented relative to the larger field. Also, to account for a contest scoring difference, CET entrants who have missed three or four contests out of five are scored with a number-letter combination, what this means is that their point total fits into the gap after that rank in the main CET scoring table (reserved for those who have entered three to five contests now). Usually there will only be one score in that gap so the letter A will be used but if there's two or three you may see B and C used too. These are not CET contest official ranks, once you're out of the main table there's no way back in and you're ranked with others of your kind so to speak. The EWP ranks are entry-number-blind so you can stay in the main table, however, the same protocol is used for average error ranking. So there, I have gone to the same system. Unranked average error stats are compared using a number-letter system. The CET contest works in the opposite direction, while total scores are separated for infrequent entrants, their average error is blended in with all contest entrants. But to provide a more consistent reporting on ranks, I have adjusted the ranks from overall to same system so that some high ranked infrequent entrants are now re-ordered by a number-letter combination. Your actual rank in the CET contest (avg error category) can be worked out by either checking this table for higher-ranked (lower numbers) infrequent entrants with a letter, and assuming them to be ahead of you, or by going into the CET scoring file provided by J10 and subtracting the number in the avg error column from 112 since the points are reverse of ranks (and 111 is the total number of ranks given). If a rank is shown as 0  with a letter (zero) it means the value is better than first place among frequent entrants. Thus 0A means best of these and 0B, 0C would be second and third best of them.  

    One other note, if somebody has entered a different number of contests, and the EWP count is smaller so that the ranks are considerably different, I could treat that situation like a CET-only with the EWP rank in brackets. The number of contests entered is the number in brackets after forecaster name. If it's different, the two numbers are (CET, EWP) contests played. If the differential is not too big then I blend the ranks. 

    I have only integrated those with two contests, all scores for those with one contest can be seen in the two excel files. 

    If anyone can't follow this format, ask (private message might be best). 

     

    RANK __avg _ CET _ EWP __Forecaster ________ avg C err (rank) __ avg mm error (rank) 

    _ 01 ___ 04.5 __ 01 __ 08 ___ Reef (5) _____________ 0.70 (2) _______36.28 (8) 

    _ 02 ___ 08.0 __ 14 __ 02 ___ Feb1991Blizzard ____ 1.00 (20t) ____ 37.56 (5) 

    03 ____ 09.0 __ 09 __ ---- ______Quicksilver1989 (5) _ 0.80 (5t)  

    _ 04 ____11.5 __ 08 __ 15 __ noname_weather (5) ___0.82 (7) _______ 46.84 (28)

    _t05 ___ 12.0 __ 20 __ 04 ___ Federico (5) _________0.96 (13) ______ 35.96 (5) 

    _t05 ____ 12.0 __ 12 __ ---- ______ Summer Sun (5) ____0.98 (14t) 

    _ 07 ____15.0__ 17 __ 13 ___ Polar Gael (5) ________1.02 (23t) ______ 33.04 (1) 

    _ 08 ____15.5 __ 10 __ 21 ___ Leo97t (5) ___________1.00 (20t) ______ 55.24 (50) 

    _ 09 ____ 16.0 __ 16 __ ---- _______ damianslaw (5) ____ 0.88 (9t) 

    _ 10 ___ 17.5__ 02 __ 33 ___ Summer18 (5) _______ 0.80 (5t) ________ 47.88 (33)

    _t11 ____ 18.0 __ 18 __ ---- ______ dancerwithwings (5)_ 1.02 (23t) 

    _t11 ___ 18.0 __ 33 __ 03 ___ Bobd24 (5) __________ 1.10 (31t) _______ 39.08 (10) 

    _t13 ___ 19.0 __ 15 __ 23 ___ General Cluster (5) __ 1.20 (34t) _______ 44.88 (t21) 

    _t13 ____ 19.0 __ 19 __ ---- ______ Man Without Beard (5)_0.98 (14t) 

    _ 15 ___ 19.5 __ 29 __ 10 ___ Don (5) ______________ 1.12 (33) _______ 41.60 (13) 

    _ 16 ____21.5__ 24 __ 17 ___ Roger J Smith (5) _____1.30 ( 40) ______ 46.84 (30)

    _ 17 ____22.0__ 26 __ 18 __ davehsug (5) __________0.98 (16t) ______ 44.80 (20)  

    _t18 ____23.5 __ 22 __ 25 __ J10 (5) ________________1.06 (26t) ______ 47.58 (31) 

    _t18 ____23.5 __ 31 __ 16 __ February1978 ________1.20 (34t) ______ 46.40 (27)

    _t18 ____23.5__ 05 __ 42 __ 2010 cold (4) __________0.64 (1) ________ 44.23 (17)

    _t21 ____24.0__ 41 __ 03 __ Relativistic (5) _________1.20 (34t) ______ 34.24 (3)

    _t21 ____24.0 __11 __ 37 __ MrMaunder (5) ________0.94 (12) _______ 50.48 (38)  

    _ 23 ____24.5__ 13 __ 36 __ BornFromTheVoid (5) ____1.10 (31t) _______47.60 (32) 

    _t24 ____25.0 __21 __ 29 __ Summer Blizzard (5) __ 1.00 (20t) _______45.60 (23) 

    _t24 ____25.0__ 03 __ 47 __ weather-history (5) ____0.74 (3) _________53.40 (44)

    _t24 ____25.0__ 28 __ 22 __ stewfox (5) ____________1.38 (43) ________52.84 (43) 

    _t24 ____25.0__ 44 __ 06 __ The PIT (5) _____________1.34 (42) _______ 39.84 (11)

    _t24 ____ 25.0 __ 25 __ ---- _____ Mark Bayley (4) _______0.90 (11t)

    _ 29 ___ 27.0 __ 43 __ 11 ___ Jeff C (5) ______________1.40 (44t) ________42.08 (14) 

    _ 30 ____27.5__ 23 __ 32 __ pegg24 (5) _____________0.98 (14t) _______ 50.40 (37) 

    _t31 ____28.0__ 32 __ 24 __ SteveB (5) _____________1.06 (26t) ________ 48.00 (34) 

    _t31 ____28.0__ 06 __ 50 __ Stargazer (5) __________ 0.84 (8) __________ 58.48 (53)

    _t31 ____28.0__ 07 __ 49 __ DR(S)NO (5) ___________ 0.88 (9t) _________ 54.88 (48)

    _ 34 ____29.0__ 27 __ 31 __ Weather26 (5) _________ 1.28 (38t) _______ 50.32 (36)

    _ 35 ____29.0__ 30 __ 28 __ Mulzy (5) ______________ 1.04 (25) _________46.20 (26) 

    _ 36 ____30.0__ 48 __ 12 __ Godber 1 (5) __________ 1.44 (48t) ________ 43.68 (15)

    _t37 ____31.0__ 53 __ 09 __ Blast from the Past (5) _1.28 (38t) ________ 44.88 (21t)

    _t37 ____31.0 __ 62 __ 20 __ jonboy (5) _____________1.82 (59t) ________ 44.60 (19)

    _t39 ____31.5__ 04 __ 59 __ timmytour (5) _________ 0.78 (4) __________ 60.48 (56)

    _t39 ____31.5__ 36 __ 27 __ Kirkcaldy Weather (5) ____ 1.52 (51) ________ 53.96 (46) 

    _ 41 ____ 32.0__ 59 __ 05 __ snowray (5) ___________ 1.74 (58) ________ 36.20 (7)

    _ 42 ____33.5 __ 37 __ 30 __ Norrance (4) __________ 1.23 (36) ________ 53.93 (46)

    _t43 ____34.0 __ 49 __ 19 __ seaside60 (5) __________1.56 (52) ________ 46.00 (25)

    _t43 ____34.0__ 34 __ 34 __ DiagonalRedLine (5) ___ 1.32 (41) ________ 57.44 (52) 

    _t43A ___34.0 __ 34A _(70)_ Nigerian Prince (2,1)___ 0.90 (11A) _______ 4.30 (0A) __ EWP rank not avg  

    _t45 ____35.0__ 69 __ 01 __ virtualsphere (5) _______ 2.64 (69) ________ 32.16 (2)

    _t45 _____35.0 __ 35 __ ---- _____ Froze were the days (3) __ 0.90 (11t) 

    _ 47 ____37.0 __ 60 __ 14 __ Midlands Ice Age (5) ___ 2.10 (62t) ________ 41.38 (12)

    _t48 ____ 38.0 __ 50 __ 26 __ Stationary Front (5) ___ 1.88 (61) _________ 47.08 (29)

    _t48 ____ 38.0__ 38 __ 38 __ Frigid (4) _______________ 1.40 (44t) ________ 43.90 (16)

    _t50 ____ 40.0__ 39 __ 41 __ sundog (4) _____________ 1.50 (50) _________ 60.33 (55)

    _t50 _____ 40.0 __ 40 __ ---- ______ Duncan McAlister (5) _1.82 (59t)  

    _ 51A___ 40.5 __ 36A__ 45 __ coldest winter (2) ______ 0.94 (12A) _______ 20.6 (0C)

    _ 52 _____ 42.0 __ 42 __ ---- ______ Typhoon John (5) _____ 1.40 (44t) 

    _ 53 ____ 43.0__ 47 __ 39 __ shillitocettwo (5) _______ 2.16 (64t) ________ 52.40 (41) 

    _ 53A___ 44.0__ 30A__58 __ cawsand (2) ____________ 0.60 (01A) _______ 43.45 (14A)

    _ 54 ____ 47.0__ 51 __ 43 __ moffat (3) ______________ 1.27 (37) _________37.03 (9)

    _ 54A ___ 48.0 _ 48A _(88) _ BARRY (2,1) ____________ 0.94 (12A) ________ 13.70 (0B)

    _ 55 ____ 49.0__ 54 __ 44 __ syed2878 (5) ___________ 1.66 (55) _________ 54.38 (47)

    _ 56 ____ 50.0__ 65 __ 35 __ emmett garland (5) ____ 2.16 (64t) ________ 49.80 (35)

    _ 57 ____ 50.5__ 55 __ 46 __ Earthshine (4) __________ 1.40 (44t) ________45.90 (24) 

    _ 58 ____ 51.5__ 63 __ 40 __ Walsall Wood Snow (5) _ 2.34 (68) ________ 51.22 (39)

    _ 59 ____ 53.5__ 52 __ 55 __ I Rem Atl 252 (5) ________1.44 (48t) ________62.36 (58)

    _ 60 ____ 54.0__ 45 __ 63 __ CheesepuffScott (3) ____1.07 (28t) ________55.17 (49)

    _ 61 ____ 55.0__ 56 __ 54 __ Joneseye (3) ____________1.07 (28t) ________34.53 (4)

    _ 62 ____ 57.0__ 61 __ 53 __ Dog Toffee (3) __________1.60 (53) ________ 56.90 (51)

    _ 63 _____ 58.0 __ 58 __ ---- ________ Kentish Man (5) _____ 1.68 (56t) 

    _ 63A ___ 59.5 _ 71A _ 48 __ cymro (2) _______________ 3.70 (70B) ______ 26.75 (0D)

    _t64 ____ 61.0 __ 66 __ 56 __ Neil N (5) _______________ 2.10 (62t) _______ 59.00 (54)

    _t64 ____ 61.0 __ 70 __ 52 __ booferking (3) __________2.77 (70) ________ 52.90 (42)

    _64A ____61.0 __ 53A_ 69 __ B87 (2) _________________ 1.50 (50A) ______ 77.70 (61B)

    _ 66 ____ 61.5__ 72 __ 51 __ Sleety (3) _______________ 4.23 (71) ________ 44.37 (18)

    _t67 ____ 62.5__ 64 __ 61 __ moorlander (3) _________ 1.63 (54) ________ 71.03 (61)

    _t67 ____ 62.5__ 68 __ 57 __ Let It Snow! (3) _________ 2.23 (66t) ________ 62.23 (57)

    _ 69 ____ 63.5 __ 67 __ 60 __ DAVID SNOW (3) _______ 2.23 (66t) ________ 52.23 (40)

    _ 70 ____ 64.0 __ 46 __ 82 __ prolongedSnowLover (3) 1.07 (28t) _______ 70.70 (60)

    _ 71 ____ 65.5 __ 57 __ 74 __ daniel* (5) ______________ 1.68 (56t) _______ 63.00 (59)

    _ 71A ___ 66.0 _ 70C _ 62 __ Paul T (2) ________________ 3.60 (70A) ______ 46.75 (27A)

    _ 72 ____ 68.5 __ 71 __ 66 __ Lettucing Gutted (5,4) ___ 6.14 (72) _______203.26 (62)

    _t73A ___70.0A _70A _(97) _ Captain Shortwave (2) ___ 2.70 (69A) ______ 76.75 (61A)

    _t73A ___70.0A _70B _(96) _ SMU (2,1) ________________ 2.20 (65A) _____ 107.20 (61C)

    ==========================================================

    The rankings are cut off after 2 entries, those with one entry in either or both contests

    can see their rankings in the excel files for the contests. 

    The highest ranked one-entry CET forecasts had zero errors and have similar points to 20th place.

    The highest ranked one-entry EWP forecast is from January, john88b was ranked first and error was 5.8 mm.

    ... now in 64th place in the EWP contest. 

    note: minor edit on May 5 to adjust EWP avg err ranks, most are unchanged. 

     

    Edited by Roger J Smith
    • Thanks 1
    Link to comment
    Share on other sites

    Posted
  • Location: Woodland Garden, South Yorks
  • Weather Preferences: Sunny weather, thunderstorms
  • Location: Woodland Garden, South Yorks

    This month is like an unfolding nightmare scenario in waking life.

    Link to comment
    Share on other sites

    Posted
  • Location: Rossland BC Canada
  • Location: Rossland BC Canada

    Report on the CET and EWP contest performances of consensus and recent normals

    For the CET contest, we do not officially score consensus or the two recent normals. (In the EWP contest they are tracked). So here's an update which shows how our consensus (median rather than mean) and the two normals have been doing. 1991-2020 played December as 1990-2019. All I can report here is their rank in each contest, as the other scoring parameters are a bit too complex for me to recreate after the fact. The rank shown is equivalent to the highest scoring forecaster at that temperature level, some others would have ranked lower (higher rank numbers). 

    The number in brackets beside the rank is the number of forecasts submitted. 

     

    Robotic forecaster ____ DEC rank ___ JAN rank ___ FEB rank ___ MAR rank ___ APR rank ______ average rank

    Consensus ____________ 27th (97) ___ 16th (78) ___38th (80) ___ 22nd (62) ____ 30th (62) ______ 27th (76)

    1981-2010 normal ____ 27th (97) ____51st (78) ___ 5th (80) ____ 36th (62) ____ 37th (62) ______ 31st (76) 

    1991-2020 normal* ___ 4th (97) _____56th (78) ___ 3rd (80)____ 33rd (62) ____ 48th (62) ______ 29th (76)

    Consensus error _______ --0.4 ________ --0.3 _______ --2.1 ______ --0.3 _________ +1.9 _______ avg abs 1.0

    1981-2010 error _______ --0.4 _________ +1.3 _______ --0.7 ______ --0.6 _________ +2.1 _______ avg abs 1.0

    1991-2020 error _______ --0.1 _________ +1.6 _______ --0.2 ______ --0.5 _________ +2.6 _______ avg abs 1.0

    mean bias of consensus is --0.2, mean bias of 1981-2010 is +0.3, mean bias of 1991-2020 is +0.7

    ANALYSIS: The scoring performance of all three robotic forecasters is about the same, normals did better in Feb and our consensus did better in Jan, there was little to choose in the other three months and in April they all did quite poorly. 

    While this is not a complete scoring analysis, forecasters with average errors of 1.0 are ranked around 20th to 30th in the CET contest and as seen in the combined scoring above, there are quite a few in that range. So neither consensus nor either recent normal is a particularly helpful guide to forecasters although you could do worse than to choose any of them as your model. Probably a forecast 0.2 below the old 1981-2010 normals would be doing best of any robotic type forecasts to date. 

    It should be noted that this analysis by ranks will tend to present the robots in a slightly better frame than a randomly selected forecaster because they have no rank drops for their duplicate entries, in other words, they are always first in at their levels. If they were acting at random in that regard their average ranks would probably be 2-3 positions down (higher rank numbers). 

     

    In the EWP contest where these three are tracked, they currently stand around 7th place for the more recent of the two normals, and 9th for both our consensus and the 1981-2010 averages, so in that contest, they are doing a bit better. They have varied from monthly ranks near the middle of the pack to a first place tie for consensus in February. 

    Link to comment
    Share on other sites

    Posted
  • Location: Rossland BC Canada
  • Location: Rossland BC Canada

    The EWP has settled at 13.8 mm in the tables. That remains 10th driest April on record, just ahead of 1854 at 14.1 mm. 

    This slight variation from the 12.5 mm used in the scoring file will make almost no difference, your average error if you entered April will be reduced by 0.26 mm. Rather than posting again, I will go back and edit the posts with average error shown, but that may take a few hours. The final amount does not translate into any points differential for anyone. Only Polar Gael escaped the full chop on average error percentage where you can in most months gain a few fractions of a point by coming in within 50% of the leader. The percentages increased so quickly that this was not the case this time around. 

    The top three in April were Polar Gael, Weather26 and Roger J Smith; the annual rankings were posted and also you can see them in the combined standings two posts back although not in order. 

    • Like 1
    • Thanks 1
    Link to comment
    Share on other sites

    Posted
  • Location: Edinburgh (previously Chelmsford and Birmingham)
  • Weather Preferences: Unseasonably cold weather (at all times of year), wind, and thunderstorms.
  • Location: Edinburgh (previously Chelmsford and Birmingham)
    On 08/04/2021 at 16:52, Relativistic said:

    Recent first halves (1st-15th) of April to beat are:

    2013: 5.8C
    2008: 6.7C
    2000: 5.6C
    1998: 6.6C
    1996: 6.5C
    1994: 5.8C
    1990: 6.7C
    1989: 6.5C
    1986: 4.4C
    1984: 5.6C
    1983: 5.6C
    1978: 5.4C
    1977: 6.2C
    1975: 5.6C
    1973: 6.3C
    1970: 4.7C

    We stand a chance of beating all those since 1986, although that's far from nailed on. A slim possibility we end up third coldest since 1970.

    Forgot to check this at the end of the month. After corrections we ended up with 5.3C, so did manage third coldest since 1970.

    • Like 1
    • Thanks 1
    Link to comment
    Share on other sites

    Create an account or sign in to comment

    You need to be a member in order to leave a comment

    Create an account

    Sign up for a new account in our community. It's easy!

    Register a new account

    Sign in

    Already have an account? Sign in here.

    Sign In Now
     Share

    • Recently Browsing   0 members

      No registered users viewing this page.

    ×
    ×
    • Create New...