October 2020 C.E.T. and EWP forecast contests

[Stationary Front]

Looks like 10.5 is the final figure.

[damianslaw]

Shock hold the front page, a below average month compared to 61-90 mean, only just by 0.1 degrees, but below it is.. could have been a notably cooler than average month had it not been for the last week or so upturn. The first two thirds were notably below par.

[Roger J Smith]

EWP update _ 154 mm to 30th, est 8-10 mm for 31st, will have the provisional value tomorrow, expect mid 160s.

Steve Murr had CET right on, will be around 3rd or 4th in EWP (150 mm), for one of two most likely best combined forecasts. February 1978 may have a slightly better combined rank due to order of entry in CET (10.5, 138 mm).

These eight forecasters all said 10.5, with order of entry (the number in brackets) to determine ranking for CET scoring:

Godber1 (16), Man with Beard (28), sundog (42), February1978 (50), davehsug (54), Summer Sun (L1-3), Steve Murr (L1-4), DR(S)NO (L1-6). 

These seven were 0.1 out (all CET scoring to be confirmed by J10) ...

NeilN 10.6 (19), General Cluster 10.4 (20), CheesepuffScott 10.4 (23), Stationary Front 10.4 (27), The PIT 10.6 (36), Don 10.6 (45), DAVID SNOW 10.4 (52).

Five more forecasts were only 0.2 out; a total of 20 then were within 0.2 this month. 

[Walsall Wood Snow]

At 10.5c October 2020 shares the same mean CET with October's:

1680

1714

1715

1837

1903

1927

1930

1938

1983

2004

 

 

 

 

 

 

 

 

 

[Weather-history]

Interesting how similiar the last 4 months have been to those in 2004. The summer was similar to that in 2004. September 2004 was one degree warmer than September 2020 but the rainfalls are similiar. Both Septembers are about within 5mm of each other and both Octobers rainfall total are going to reasonably close.

[Walsall Wood Snow]

8 minutes ago, Weather-history said:

Interesting how similiar the last 4 months have been to those in 2004. The summer was similar to that in 2004. September 2004 was one degree warmer than September 2020 but the rainfalls are similiar. Both Septembers are about within 5mm of each other and both Octobers rainfall total are going to reasonably close.

Hopefully we get a white Christmas like that year too .

[Froze were the Days]

1 hour ago, damianslaw said:

Shock hold the front page, a below average month compared to 61-90 mean, only just by 0.1 degrees, but below it is.. could have been a notably cooler than average month had it not been for the last week or so upturn. The first two thirds were notably below par.

But this is a theme when we do manage to get a chilly month (fractionally under)...it's been a long time since we've had a month considerably under CET for the 1961-90 series!

[DAVID SNOW]

On 25/10/2020 at 21:03, DAVID SNOW said:

Forward from this, 10.5 final number after corrections.

As promised @Godber 1.

Well done you and the others.

[Roger J Smith]

Just barely, the month joined the list of Octobers that warmed from 16th value of CET (10.4) to final result. 

The downward adjustment was 0.2 (from 10.7) although only 0.14 when two decimals used (10.68 to 10.54). 

[The PIT]

Sunny Sheffield finished at 9.8C -0.4C, Rainfall 104.4mm 126.1% of monthly average.

[J10]

Downloads

October 20 Summary.pdf

October 20 CET.xlsx

Monthly

There were eight players who got 10.5c spot on.

Godber 1, Man With Beard, sundog, February1978, davehsug, Summer Sun, Steve Murr and DR(S)N

This month 39 were within 0.5c ( compared to 30 players last month)

Seasonal

A very high standard this season.

Man With Beard has got both temp scores correct this season.

While both sundog and DR(S)NO, have got 1 correct, and their other month only 0.1c out.

Overall

Some changes this month in the Top 5,
mb018538 stays in the lead.
Stationary Front (2nd from 4th)
Man With Beard (joint 3rd from 5th)
Don (joint 3rd from 3rd)
Dancerwithwings (5th from 2nd)

Edited November 1, 2020 by J10

[damianslaw]

4 hours ago, Froze were the Days said:

But this is a theme when we do manage to get a chilly month (fractionally under)...it's been a long time since we've had a month considerably under CET for the 1961-90 series!

Yes, not sure last time we ended up 1 degree or below the 61-90 mean, more likely to have been a summer month.

[Relativistic]

8 hours ago, Froze were the Days said:

But this is a theme when we do manage to get a chilly month (fractionally under)...it's been a long time since we've had a month considerably under CET for the 1961-90 series!

The period 2009-2013 was the last that seemed to regularly manage it. Our best chances are the Winter months.

[Roger J Smith]

Looks like last 1.0 below 1961-90 normals was in Nov 2016, and before that in March 2013 which was 3.0 below. Going back to 2009, these are the listed departures (to two decimals) in the tables for each year of "final CET values" ... showing all the months that were below or equal to 1961-90 in bold type and any small positives that round to 0.0 in italics. (2009 and 2010 were being compared to 1961-90 evidently, all same base for differentials as 2011).

YEAR __ Jan _ Feb __ Mar _ Apr __ May _ Jun __ Jul _ Aug _ Sep _ Oct _ Nov _ Dec

2009 __ -0.78 _+0.36 _+1.34 _+2.09 _+0.90 _+0.63_+0.10_+0.83_+0.65_+0.94_+2.11_-1.55 

2010 __ -2.41 _-0.98 _+0.45 _+0.94 _--0.50 _+1.06_+1.01_--0.51_+0.19_--0.28_--1.37_-5.34 

2011 __ -0.11 _+2.60 _+1.04 _+3.93 _+1.01 _-0.38_--0.83_-0.38_+1.50_+1.99_+3.02_+1.32 

2012 __+1.64 _+0.05 _+2.59 _--0.67 _+0.54 _--0.63_--0.53_+0.85_-0.60_-0.96_+0.30_+0.14 

2013 __ -0.35 _-0.57 _--2.98 _--0.42_--0.75_ --0.56_+2.29_ +0.11_+0.06_+1.90_-0.38_+1.69 

2014 __ +1.90 _+2.41 _+1.95 _+2.34_+1.09 _+0.99 _+1.67_-0.82 _+1.48_+1.86_+2.03_+0.51 

2015 __+0.60 _+0.27 _+0.71 _+1.14_--0.40 _--0.18 _--0.14_+0.14_--0.95_+0.38_+2.95_+5.03 

2016 __+1.64 _+1.14 _+0.09 _--0.43 _+1.37_+1.08_+0.81_+1.25 _+2.43 _+0.27_--1.00_+1.33 

2017 __ +0.14 _+2.33 _+2.99 _+0.98_+2.05_+1.88_+0.72 _-0.16_--0.07 _+1.73_ +0.27_+0.10 

2018 __ +1.43 _--0.83 _--0.75 _+1.95_+2.09 _+1.94 _+3.08_+0.87_+0.12_ 0.00_ +1.77_+2.24  

2019 __ +0.14 _+2.98 _+2.14 _+1.18_--0.01_+0.09_+1.40 _ +1.29_+0.68_--0.64_--0.31_+1.15  

2020 __ +2.58 _+2.54 _+1.02 _+2.53_+1.33 _ +1.16_--0.32_+1.80_+0.35_--0.08 

________________________________________________________________

Out of 142 months in this period, 41 were below the 1961-90 normals, and four more were equal. 

The three largest departures were Dec 2010, Mar 2013 and Jan 2010. 

 

Edited November 2, 2020 by Roger J Smith

[North-Easterly Blast]

16 hours ago, Weather-history said:

Interesting how similiar the last 4 months have been to those in 2004. The summer was similar to that in 2004. September 2004 was one degree warmer than September 2020 but the rainfalls are similiar. Both Septembers are about within 5mm of each other and both Octobers rainfall total are going to reasonably close.

Yes; summer 2020 had CETs in all three summer months identical or at least very close to summer 2004.  May 2020 also wasn't all that much warmer than in 2004.  October 2020 was also very similar to 2004.  

[North-Easterly Blast]

I actually calculate that October 2020 had a warmer second half than the first half (First half: 10.48, Second half: 10.58).  It is not often that the second half of October is warmer than the first half - the last occasion was I believe in 2014, and before that in 2009; so last month was an anomaly in seeing a warmer second half of October than the first half.

[Roger J Smith]

The EWP provisional end of month total is 161 mm. I think rather than doing two sets of revisions when they come up with a more precise value on the 5th, I will just wait until then to show the monthly and annual scoring, as I don't want to go through it twice (when it's this high, the value we see on the 5th could be several mm different from this one). Just for the record though, 161 would mean that shillitocettwo has the closest forecast at 156, followed by daniel* at 170, Steve Murr at 150, and Mr_Moreau (first time entrant) at 174 mm. If they happen to hit us with a final value above 163 then I would be changing the file twice in opposite directions. So let's wait and see what they say on the 5th. The scoring that I posted about a week ago is going to be very close to the final results in any case. 

Here's the original list amended for the slightly different actual result. Will confirm on 5th from the final value posted by Hadley EWP.

(edit on 5th _ List edited slightly to conform to 163.8 mm, the actual posted value ... this moves daniel* back to the full ten points high score and provides a few other minor changes in ranking therefore scoring also). 

(note, I have added all the relevant annual updates after all, including some pro-rated comparisons for those who haven't entered all the months, just for interest -- official contest ranks are the annual totals).

(( OCTOBER SCORES )) for 163.8 mm

 

01 _ daniel* ________ 170.0 ____ 10.00 (2nd scoring interval)

02 _ Mr_Moreau ____174.0 _____ 9.62 (3rd scoring interval)

03 _ shillitocettwo___156.0 _____ 9.61 (9.81 raw score - 0.2 late)

04 _ 2010 cold ______141.8 _____9.24 (5th scoring interval)

05 _ Steve Murr ____ 150.0 _____9.23 (9.43 raw score - 0.2 late)

06 _ mb018538 _____140.0 ____ 9.06

07 _ February1978__ 138.0 ____ 8.87

08 _ Mulzy __________ 135.0 ____ 8.68

09 _ DR(S)NO _______ 135.0 ____ 8.40 (0.28 deduction 2nd entry, 1d late)

10 _ Roger J Smith ___133.3 ___ 8.30

11 _ Relativistic _____ 130.0 ___ 8.11

12 _ Jonboy _________ 128.0 ___ 7.92

13 _ Lettucing Gutted 200.0 ___ 7.73

14 _ General Cluster_125.0 ____ 7.54

15 _ syed2878 ______ 125.0 ____ 7.46 (0.08 deduction 2nd entry)

16 _ J10 _____________ 125.0 ____ 7.18 (0.36 deduction, 3rd entry, 1d late)

17 _ Stationary Front_122.0 ___ 6.98

18 _ Bobd29 _________121.7 ___ 6.79

19 _ Don _____________120.0 ___ 6.60

20 _ I Rem Atl 252 ___ 210.0 ___ 6.41

21 _ Summer Shower 113.0 ___ 6.23

22_ Blast from the Past 112.0 _ 6.04

23 _ Earthshine ______ 111.0 ___ 5.85

24 _ Emmett Garland _111.0 ___ 5.77 (2nd entry deduction 0.08)

25 _ Jeff C _____________110.0 ___ 5.47

26 _ davehsug ________ 110.0 ___ 5.39 (2nd entry deduction 0.08)

27 _ virtualsphere ____ 107.0 ___ 5.10

28 _ CheesepuffScott _ 106.9 ___ 4.91

29 _ Mr Maunder ______105.0 ___ 4.72

30 _ Born from the Void 105.0 __ 4.64 (2nd entry deduction 0.08)

(31 to 54 had lower forecast amounts, generally points drop 0.2 each value except 0.08 per later duplicate)

 

UPDATED ANNUAL SCORING from October scores

^ missed month(s) ... rank in brackets if average score applied to 11 months as per most entrants

(note, this adjusted rank assumes only the one forecaster has a boosted score, if all were boosted, most would

come out a few ranks lower than shown -- what this tells you is your average forecast performance, but the

contest scoring is partly based on participation so that actual totals are relevant too, no telling how anyone 

might actually have done in the months they didn't compete).

01 _ Twilight ____________75.38

02 _ Godber.1 __________72.81

03 _ mb018538 ________ 71.36

04 _ Mulzy _____________ 71.10

05 _ Reef _______________ 69.37

06 _ Blast from the Past_ 68.66

07 _ Jeff C ______________ 66.83

08 _ Bobd29 ___________ 66.69

09 _ Emmett Garland __ 65.25

10 _ Don _______________ 64.88

11 _ DR(S)NO ___________63.44

12 _ General Cluster ___ 62.12

13 _ Born From the Void 61.83

14 _ Jonboy ____________ 61.71^ (7)

15 _ Roger J Smith _____ 61.32

16 _ daniel* ___________ 60.65

17 _ Stationary Front __ 60.45

18 _ Dog Toffee ________59.86

19 _ Weather26 ________58.68

20 _ J10 ________________ 58.10

21 _ DAVID SNOW _____ 57.11

22 _ Virtualsphere _____ 56.81

23 _ February1978 _____55.62^ (17)

24 _ Steve B ____________55.30

25 _ Feb1991Blizzard __ 54.90

26 _ snowray __________ 54.83

27 _ The PIT ___________ 53.90

28 _ davehsug _________ 53.29

29 _ Midlands Ice Age __51.68

30 _ syed2878 _________ 51.40^ (22)

31 _ Mr Maunder ______ 50.53

32 _ Relativistic ________ 49.77^ (26)

33 _ brmbrmcar _______ 49.52^ (26)

34 _ CheesepuffScott __ 48.49

35 _ I rem Atl 252 ______ 48.32^ (29)

35 _ seaside60 _________ 46.94

37 _ Timmytour ________ 44.91

38 _ Neil N _____________ 44.27

39 _ Polar Gael _________ 43.56^ (35)

40 _ B87 ________________43.30^^ (29)

41 _ Thundery Wintry Sh 43.09^^^^ (7)

42 _ weather-history ____42.88

43 _ Stargazer ___________41.42^^ (32)

44 _ Earthshine _________ 41.33^^^^ (10)

45 _ prolonged SnowLover 40.66 ^^ (33)

46 _ Shillitocettwo _______ 38.36^^^ (28)

47 _ Norrance ___________ 38.21

48 _ Kirkcaldy Weather __ 36.69^ (46)

49 _ Diagonal Red Line __ 36.69 ^^^ (31)

some ranks now missing, only Oct participants are listed

52 _ Summer Blizzard ___ 30.36 ^^^^ (36)

54 _ Summer Shower ___ 29.51 (4 entered, avg 7.22, would total 79.42 1st)

56 _ 2010cold ___________ 28.74 (5 entered, avg 5.72, would total 62.92 12th)

68 _ Steve Murr __________10.76 (2 entered, avg 5.38, would total 59.18 19th)

73 _ Lettucing Gutted ____10.12 

76 _ Mr Moreau _________ 9.62 (1 entry, equivalent to 105.82 points 1st)

77 _ Thundershine _______ 9.57 (5 entered, would be 59th at same rate for 11)

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

Some good scoring from Summer Shower and 2010cold, and Mr Moreau's lone effort. We'll see if they become regular

participants in the next contest year how things go (Mr Moreau a no show in Nov, our fourth contest year one and out

top three scorer). 

 

 

Edited November 5, 2020 by Roger J Smith

[Roger J Smith]

Updated version of earlier list of Octobers with warmer second halves ... 2020 did barely scrape onto the list ...

As to Octobers with warmer second halves than 1st to 16th, this is the complete list of 33 cases (from 1772) ... shows the final CET against the 1-16 CET ... in order of differentials and then chronological within same set. There were also 10 cases of no change in CET from 16th to 31st, the most recent 1980 (9.0) and 2004 (10.5), 2005 (13.1). ... possibly also 2015 which in my data dropped from 11.05 to 10.99. 

Since 43 years had either an increase or no change, the other 205 in the data set dropped by some amount, 1859 managed a full 4 deg drop from 13.6 to 9.6 at the end. 

1849 __ 9.2 (7.6) __ +1.6 ______ 1897 __ 9.9 (9.5) __ +0.4

1888 __ 7.9 (6.4) __ +1.5  ______ 1898 _ 11.3 (10.9) _+0.4

1867 __ 9.3 (8.2) __ +1.1  ______ 1927 _ 10.5 (10.1) _+0.4

1803 __ 9.3 (8.5) __ +0.8  ______ 1872 __ 8.4 (8.1) __ +0.3

1809 _ 10.2 (9.4) __+0.8  ______ 1884 __ 9.4 (9.1) __ +0.3

1952 __ 8.8 (8.1) __ +0.7  ______ 1988 _ 10.4 (10.1) _+0.3

2014 _ 12.5 (11.8) _+0.7  ______ 1779 _ 10.9 (10.7) _+0.2

1833 _ 10.1 (9.5) __+0.6  ______ 1826 _ 11.1 (10.9) _+0.2

1860 __ 9.8 (9.2) __ +0.6  ______ 1830 _ 10.4 (10.2) _+0.2

1925 _ 10.4 (9.8) __ +0.6  ______ 1899 __ 8.8 (8.6) __ +0.2

1838 __ 9.8 (9.3) __ +0.5  ______ 1936 __ 9.4 (9.2) __ +0.2

1853 _ 10.1 (9.6) __ +0.5  ______ 2009 _ 11.6 (11.4) _+0.2

1871 __ 9.8 (9.3) __ +0.5  ______ 1792 __ 8.8 (8.7) __ +0.1

1904 __ 9.7 (9.2) __ +0.5  ______ 1789 __ 8.6 (8.5) __ +0.1

1923 __ 9.7 (9.2) __ +0.5  ______ 1864 __ 9.8 (9.7) __ +0.1

1977 _ 11.8 (11.3) _+0.5  ______ 1877 __ 9.3 (9.2) __ +0.1

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1989 _ 11.7 (11.6) _ +0.1

--- --- --- --- --- --- --- --- --- --- --- -- 2020 _ 10.5 (10.4) _ +0.1

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

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

Has to be said that this is not the most encouraging list of analogues for a winter forecast. 

Note _ have edited in the 2020 result which barely makes this list. 

[Froze were the Days]

10 hours ago, Relativistic said:

Our best chances are the Winter months.

Going by Roger's informative post it 'use' to be the winter month's...not a lot has happened in the last 6 years or so other than the Feb 2018 classic SSW.

[NeilN]

Again so close...fourth overall for this season is pretty good going mind :)

 

[summer blizzard]

Pretty close on the actual outcome but little movement higher in the competition given how closes to the consensus I was. 

[Roger J Smith]

10 hours ago, Froze were the Days said:

Going by Roger's informative post it 'use' to be the winter month's...not a lot has happened in the last 6 years or so other than the Feb 2018 classic SSW.

I was surprised there were that many (41/142), but note that in the first five of the twelve years, the average (2009-13) was 5 months a year falling below the old 61-90 averages, and the number for the past seven (2014-20) averages only two and a bit. So the appearance of them has been reduced by more than half since that run in first half of 2013. I don't think that's all down to climate change, perhaps a cyclical thing that may return to higher levels like we saw earlier. 

[Froze were the Days]

1 hour ago, Roger J Smith said:

I was surprised there were that many (41/142), but note that in the first five of the twelve years, the average (2009-13) was 5 months a year falling below the old 61-90 averages, and the number for the past seven (2014-20) averages only two and a bit. So the appearance of them has been reduced by more than half since that run in first half of 2013. I don't think that's all down to climate change, perhaps a cyclical thing that may return to higher levels like we saw earlier. 

Also of note going by that CET series it's well over due a below December or January (or both) so if you believe in sequences or them being broken (if there is such a thing in climatology) maybe it's about time for that hoodoo to be broken. 

[Roger J Smith]

My avatar has hoodoos in it (those rock pinnacles). So maybe that's a good omen. 

Not sure if there's such a thing in climatology but either the string will be broken or there will never be another cold month (now I am channelling the long lost Stratos Ferric). 

[Roger J Smith]

The EWP final posting is in at 163.8 mm. There are small changes in the monthly and annual scoring, near the top, and two ties were broken --  the previous tie for 16th is broken with Lettucing Gutted now at 13th and Stationary Front at 17th with previous 13th to 15th dropping down one, and for 21st, I Remember Atlantic 252 has moved up to 20th with Blast from the Past now 22nd and Summer Shower dropping from 20th to 21st. Those who who dropped down lost 0.19 points in both the October and annual lists, two near the top gained 0.19, and the two who moved up from the tied scores gained multiples of 0.19. I have edited these changes into the post. 

Just for general info, I have decided that the best way to adjust monthly ranks for lateness in the CET would be to consider one day late the equivalent of dropping down one rank (one rank number higher), two days late two ranks, etc. See recent posts in the November thread for discussions related to this subject.

The CET scoring formulae in use only penalize rankings for seasonal and annual points totals and that is their average effect there, so it makes sense (and conforms to roughly the same outcome in EWP where monthly ranks are directly affected by late penalties) -- this is only significant at all in terms of working out the rankings for "best combined forecast" as I do each month.  I thnk the math on this actually leans closer to a two rank CET adjustment in the month where the late penalty occurs, but CET scoring has fixed intervals whereas EWP scoring jogs slightly in favour of equalizing scores for same forecast values (you don't drop the full interval for being second in, etc). That complicates the average rank differential for late penalties in EWP, and the combination of being late and a duplicate forecast value seems to be that it prevents a rank drop because the softer hit for duplicate entry provides a higher platform for the late penalty -- I've noticed in scoring that it takes a two day late penalty in almost all cases to jog a rank out of sequence for a duplicate entry EWP forecast. There is also the difference that you can be tied for ranks in EWP if your forecast has same error, different sign (as an example +20 = -20 mm for scoring, regardless of order of entry) There is no similar effect in CET scoring because nobody can be tied or dropped less than one interval in that method. So the late penalties will always have the same relative effect every month for CET regardless of position, unless you were already dead last and had a late penalty too.

Now in the particular case we have before us, eight people predicted 10.5, they appear to rank 1 to 8, but the last three were all late one day. That means in my approximation method, 6th (Summer Sun) remains 6th overall, 7th (Steve Murr) remains 7th, 8th (DRSNO) drops to 9th while 9th in the list (NeilN) rises to 8th. Had there been only six with one late, that sixth would have gone to 7th place. These adjusted ranks are only of the limited use in the contest of applying to "best combined forecast." You'll find that if you enter every CET forecast contest late, your ranks will drop (move to a higher number) by more than just one position, probably 2-3 is more likely.

So what that means is that 7th ranked CET (Steve Murr) had 5th ranked EWP score although 4th ranked error (12 total ranking) and 4th ranked CET (February 1978) had 7th ranked EWP (both score and error), with a total 11 ranking so under the rules that's the best combined forecast although in terms of error Steve had it by a small margin.  Anything under 12 in total ranking gets you into a list of top thirty all-time combined forecasts over three years. You can find that in the excel file when I download it later today. I'll provide a short guide to the file then. 

DR(S)NO was 9th in both contests (18 total ranking) and Don was 12th in CET and 19th in EWP (31 total ranking). 

Edited November 5, 2020 by Roger J Smith