December 2020 C.E.T. and EWP forecast contests -- start of the new contest year for 2020-2021

[Summer Sun]

4.5c to the 11th

0.5c below the 61 to 90 average
0.7c below the 81 to 10 average

___________________________

Current high this month 5.1c on the 2nd
Current low this month 3.9c on the 8th & 9th

[Roger J Smith]

EWP seems headed for higher portions of our forecast table, where 125 mm was second highest prediction (the usual 200 mm from LG).

We have already reached about 57 mm and look to add 35-40 mm to 22nd, then possibly 15-20 more to 27th. 

The only suspense may be how far into the middle of the scoring LG falls, if the outcome is 125 to 140 that would be fairly far down, if 141-160 it becomes around middle of table, at 161 LG is as high as 7th place. 

The fact that it's wetter than many had expected is a byproduct of the fact that it's also milder than many foresaw, at least the projection from today to end of the GFS run is milder (around 5.5 to 6.0). We have only 11 of 97 forecasts at or above 6.0, with another five at 5.9 (looks like another chance to hit that elusive outcome). The best combined forecast that I see from these projections is that of SteveB (5.9, 125) which was the first one posted too. 

[The PIT]

Sunny Sheffield now at 4.2C -1.8C below normal. Rainfall 32mm 36.8% of the monthly average.

Onward and upward fro the rest of the week.

[Summer Sun]

4.7c to the 12th

0.2c below the 61 to 90 average
0.4c below the 81 to 10 average

___________________________

Current high this month 5.1c on the 2nd
Current low this month 3.9c on the 8th & 9th

[Roger J Smith]

Minor changes in EWP scoring procedures will make the EWP scoring even more similar to CET scoring for this year and into the future. I am not going to apply these changes retroactively but I did so as an experiment for the top five finishers last year, results were not greatly different. 

The scoring intervals will remain the same, and match the CET in that regard (except that the CET contest uses a base of 100 points and the EWP uses a base of 10 points each month). 

Forecasts that are the same value but later entries will now drop two-thirds of an interval each rather than half which was the average (for most months I was dropping them .08 with slightly variable fractions of the actual interval, after using .05 for more crowded December 2019). I will continue the practice of giving equal scores to equal errors on either side of the outcome, as this is somewhat less likely to happen in EWP than CET (because EWP has a differential from a decimal outcome that is one decimal greater than most of the forecasts, 90% of entrants use round numbers of mm). 

Late penalties are increasing from 0.2 per day to 0.3 -- this seems closer to the same impact the CET late penalties have on eventual scoring outcomes. The one difference will be that you'll see your impact on your monthly score whereas in CET scoring you don't see late penalty impacts until you get to seasonal and annual scoring. 

Bonus scoring points will now be available for forecasts on the same side of 1981-2010 normal values, .00 change will be applied to those and .06 subtracted from the guesses on the wrong side of normal. A slider will be placed in this metric, the first two guesses on each side of normal will only  lose .01, .02, .04 and .05; any guesses that match the normal will always have reductions of .03.  In the rare event of an outcome hitting the normal value or within two forecasts of it, this correction will be applied in reverse (no reduction at the normal or best forecast, smaller ones close to it expanding to .03 on either side). Then for accuracy, as I want the top score to remain 10, there will be a field-wide accuracy reduction starting with zero reduction for closest, .01 for second closest, .02 for third closest, and so on, until reaching eleventh closest forecast which would be reduced .10, as will all others less accurate. 

Finally, to allow average error to participate more in scoring, without having the same system as CET scoring of tracking it directly as a component, the same result can be achieved by toggling the scores by comparing errors to overall distribution. This will work as follows -- after all other adjustments are made to scoring as above, then your error size will be used to determine whether you qualify for .01, .02, .03, .04 or .05 scoring protections for falling within 10%, 20%, 30%, 40% or 50% of the top forecast (by differential regardless of penalties) -- for example, if the error of the top forecast is 1.5%, then all forecasts to 11.5% will lose no points, all between 11.6 and 21.5 will lose only .01 , all between 21.6 and 31.5 would lose .02, etc. ... This will maintain 10.00 as the top possible score. The only mathematical possibility for a lower top score than 10.00 comes from late penalties, I don't think it has happened in the first three years that top score was second scoring interval which would be the case if top forecast was late (I had to reduce second and fourth highest raw scores by a late penalty in October, bringing them to third and fifth). This adjustment will not be rank-related, it could apply to half the field or to one or two people depending on the distribution of errors. 

These adjustments may seem small but if you consistently make the same sorts of forecasts that should be boosted in score, over the course of a contest year they could change your ranking by 2 to 4 places. The fact that it isn't more is probably more due to random variability (everyone from time to time qualifying for large or small improvements) than any other factor. 

Here's an example of a forecast that benefits to a maximum from all these changes: 

Outcome is 60 mm, well below monthly normal value of 72.4 mm, and top score (on time forecast) is 59 mm (a 1.67% error). We'll say there were 51 forecasts, scoring interval is then 0.20. 

Second closest forecast is 61.5 mm, a 2.5% error. This loses (only) .01 for being second closest, .00 for being same side of normal, and .00 for being within ten per cent of the minimum error. Score is therefore 9.79 (in the previous system 9.80).

Seventeenth closest forecast is 73 mm and was the second entry at that value as well as being a day late. This is scored on the basis of 15th scoring interval minus duplication points, as 15th place is 14 scoring intervals, the base score at that prediction is 7.20 and then reductions are .13 rather than .20 per entry, so this one has a score of 7.20 - 0.13 - 0.30 = 6.77 (this is already 0.15 lower than in the old system of .08 drops for duplicate entries and 0.2 late penalties). For finishing lower than 11th, a reduction of 0.10 applies. Then as this is the second closest forecast to normal on the wrong side, it loses .05 (the next one and all others lose .06). Now the score is down to 6.62. The forecast error is 21.67% which is just within the second least reduction range of .02 (score now 6.60). The total reduction from the old system for this forecast is now 0.32. If in fact some others had entered 47 mm with the same errors and late penalties, but on the right side of normal, their total reductions would have been 0.27 from the old scoring system.

(A reminder that second place has dropped only 0.01 from that system).

The second score of 9.79 now has an advantage of 3.19 points compared to 2.88 in the old system of scoring (or 3.14 vs 2.88 for the 47 mm prediction).

When I list the monthly scoring I will show these adjustment factors so you can see in real time how they apply. The differentials may appear small but over time those advantages (if achieved several times) would add up to the equivalent of several annual contest ranking positions. If we all share these at random over a year then it will just be more work for me and no other changes but we'll see what happens.

There is no scoring implication for any monthly rank adjustments (beyond the lower score that forced them) and this would likely drop the "17th closest" forecast from 17th to 18th place as before, except that the drop would be further into the score spacing between ranks.  For somebody with a large late penalty and no protection from duplicate entry adjustment, the change could be 3-4 ranks instead of the current 2-3 (which would only have an impact on the best combined forecast stat). 

This should also equalize the scoring somewhat for 11 and 12 forecast entries, if the person only entering 11 times has better than average accuracy stats then they will gain those advantages more often and move up against the field of 12 entries. If they don't have better stats then it will have a similar effect to equally ranked (for error and accuracy points) 11 and 12 forecast entrants in the CET since total score is weighted roughly 2/5 of the total score. It's going to remain difficult for an 11-month entrant to win the EWP since statistically you are bound to lose 5 points to the top contenders by missing that one month. Therefore I recommend that if you are a contender that you file a forecast even in some month where you think your CET guess would be at a disadvantage from coming in late or unprepared, any score in EWP is better than no score (not necessarily true for CET contest). 

One further rule to note, no total scores can fall below zero, so in the scoring range of 0.00 to 1.00, all derived scoring changes will be re-scaled to fit the linear difference between adjusted top score in that low range, and zero. So the bottom few scores won't fall to zero and the bottom score will remain zero.

Edited December 13, 2020 by Roger J Smith

[Roger J Smith]

As a partial illustration of the changes in scoring here's how the scoring would look this month if the outcome is a little over 125 mm as expected, will set the outcome at 130 mm ... with 83 forecasts the scoring interval would be 0.122 (100/82) which is applied as 0.12 with downward jogs of .01 every 10th score. 

(edit 25 Dec _ reset to 150 mm)

(edit 29 Dec _ reset to 154 mm)

This table will show the new more complex scoring vs the old basic scoring. 

The duplicate entry drop is .08 as before because of this larger turnout. When or if the turnout drops back to 50-60 it will then be higher, but without this scoring change I probably would have applied .06 to this month's duplicate entries. 

In this scoring table, "Int score" is the scoring interval modified by duplicate entries, rank shows the deduction for forecast error ranking, acc shows the accuracy deduction, norm shows the side-of-normal deduction, and late shows the late penalty (which in this case only applies to one forecast which ranks in the top eight).  The "old system" scoring is what the 2018-19 contest scoring would have given for the same forecaster's score. You'll note that this is generally similar in the top ten and grows to a differential of near 0.2 by middle of the table. Generally the higher scores are being rewarded for accuracy in this method.

Rank __ Forecaster _____________ Fcst ___ Err ___ % err __ Int score __ rank__acc__norm__late __ Total score __ (old system)

_ 01 ___ SteveB _________________ 125.0 _ -29.0 ___18.8 _____ 10.0 __ -0.00_ -0.00_ -0.00_ -0.00 __10.00 ____ 10.00

_ 02 ___ coldest winter __________123.6 _ -30.4 ___19.8 _____ 9.88 __ -0.01_ -0.00_ -0.00_ -0.00 ___ 9.87 _____ 9.88

_ 03 ___ Fozfoster _______________120.0 _ -34.0 ___22.1 _____ 9.76 __ -0.02_ -0.00_ -0.00_ -0.00 ___ 9.74 _____ 9.76

_ 04 ___ summer shower ________115.0 _ -39.0 __ 25.2 _____ 9.64 __ -0.03_ -0.00_ -0.00_ -0.00 ___ 9.61 _____ 9.64

_ 05 ___ Don _______________ (2) _ 115.0 _ -39.0 __ 25.2 _____ 9.56 __ -0.03_ -0.00_ -0.00_ -0.00 ___ 9.53 _____ 9.58

_ 06 ___ AWD ____________________111.0 _ -43.0 __ 27.9 _____ 9.40 __ -0.05_ -0.00_ -0.00_ -0.00 ___ 9.35 _____ 9.40

_ 07 ___ davehsug __________ (2) _111.0 _ -43.0 __ 27.9 _____ 9.32 __ -0.05_ -0.00_ -0.00_ -0.00 ___ 9.27 _____ 9.34

_ 08 ___ swebby ________________ 109.0 _ -45.0 __ 29.2 _____ 9.16 __ -0.07_ -0.01_ -0.00_ -0.00 ___ 9.08 _____ 9.04

_ 09 ___ Surprised Weather _ (3) _111.0 _ -43.0 __ 27.9 _____ 9.24 __ -0.05_ -0.00_ -0.00_ -0.30 ___ 8.89 _____ 9.02

_t10 ___ Born From the Void ____ 108.0 _ -46.0 __ 29.9 _____ 8.92 __ -0.09_ -0.01_ -0.00_ -0.00 ___ 8.82 _____ 8.92

_t10 __ Lettucing Gutted ________ 200.0 _ +46.0 __ 29.9 _____8.92 _ -0.09_ -0.01 _ -0.00_ -0.00 ___ 8.82 _____ 8.92

_ 12 ___ I Remember Atl 252 ____ 105.0 _ -49.0 __ 31.8 _____ 8.68 __ -0.10_ -0.01 _ -0.00_ -0.00 ___ 8.57 _____ 8.68

_ 13 ___ --uksnow-- _________ (2) _ 105.0 _ -49.0 __ 31.8 _____ 8.60 __ -0.10_ -0.01 _ -0.00_ -0.00 ___ 8.49 _____ 8.62

_ 14 ___ Feb1991Blizzard ________ 104.0 _ -50.0 __ 32.5 _____ 8.43 __ -0.10_ -0.01_ -0.00_ -0.00 ___ 8.32 _____ 8.43

_ 15 ___ summer18 ______________ 102.0 _ -52.0 __ 33.7 _____ 8.31 __ -0.10_ -0.01_ -0.00_ -0.00 ___ 8.20 _____ 8.31

_ 16 ___ JeffC ________________ (2) _ 102.0 _ -52.0 __ 33.7 _____ 8.23 __ -0.10_ -0.01_ -0.00_ -0.00 ___ 8.12 _____ 8.25

_ 17 ___ Federico ____________ (3) _ 102.0 _ -52.0 __ 33.7 _____ 8.15 __ -0.10_ -0.01_ -0.00_ -0.00 ___ 8.04 _____ 8.19

_(18)___ 1990-2019 avg ___________101.5 _ -52.5 __ 34.0 _____ 8.01 __ -0.10_ -0.01_ -0.00_ -0.00 ___ 7.90 _____ 8.01

_ 18 ___ snowray _________________ 101.0 _ -53.0 __ 34.4 _____ 7.95 __ -0.10_ -0.01_ -0.00_ -0.00 ___ 7.84 _____ 7.95

_ 19 ___ shillitocettwo ____________ 100.0 _ -54.0 __ 35.0 _____ 7.83 __ -0.10_ -0.01_ -0.01_ -0.00 ___ 7.71 _____ 7.83

_ 20 ___ Philglossop __________ (2) _100.0 _ -54.0 __ 35.0 _____ 7.75 __ -0.10_ -0.01_ -0.01_ -0.00 ___ 7.63 _____ 7.77

_ 21 ___ General Cluster ______(3) _100.0 _ -54.0 __ 35.0 ______7.67 __ -0.10_ -0.01_ -0.01_ -0.00 ___ 7.55 _____ 7.71

_ 22 ___ St Coolio _____________(4) _100.0 _ -54.0 __ 35.0 ______7.59 __ -0.10_ -0.01_ -0.01_ -0.00 ___ 7.47 _____ 7.65

_ 23 ___ Blast From the Past __(5) _100.0 _ -54.0 __ 35.0 ______7.51 __ -0.10_ -0.01_ -0.01_ -0.00 ___ 7.39 _____ 7.59

_ 24 ___ Downburst ________________98.0 _ -56.0 __ 36.3 ______ 7.22 __ -0.10_ -0.01_ -0.02_ -0.00 ___ 7.09 _____ 7.22

_(25)___ 1981-2010 avg ____________ 97.4 _ -56.6 __ 36.5 ______ 7.16 _ -0.10_ -0.01_ -0.03_ -0.00 ___ 7.04 _____ 7.16

_ 25 ___ Leon1 _____________________ 96.0 _ -58.0 __ 37.6 ______ 7.10 _ -0.10_ -0.01_ -0.04_ -0.00 ___ 6.95 _____ 7.10

_ 26 ___ earthshine _________________95.0 _ -59.0 __ 38.3 ______ 6.98 _ -0.10_ -0.01_ -0.05_ -0.00 ___ 6.82 _____ 6.98

_ 27 ___ virtualsphere ______________ 94.0 _ -60.0 __ 38.8 ______ 6.86 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 6.68 _____ 6.86

_ 28 ___ jonboy ________________ (2) _ 94.0 _ -60.0 __ 38.8 ______ 6.78 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 6.60 _____ 6.80

_ 29 ___ Joneseye ______________ (3) _ 94.0 _ -60.0 __ 38.8 ______ 6.70 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 6.52 _____ 6.74

_ 30 ___ matt 111 ___________________92.0 _ -62.0 __ 40.2 ______ 6.50 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 6.32 _____ 6.50

_ 31 ___ Godber1 ______________ (2) _ 92.0 _ -62.0 __ 40.2 ______ 6.42 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 6.24 _____ 6.44

_ 32 ___ Reef _______________________ 91.0 _ -63.0 __ 40.8 ______ 6.25 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 6.07 _____ 6.25

_ 33 ___ bobd29 ____________________ 90.0 _ -64.0 __ 41.5 ______ 6.13 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.95 _____ 6.13

_ 34 ___ The PIT ________________ (2) _ 90.0 _ -64.0 __ 41.5 ______ 6.05 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.87 _____ 6.07

_ 35 ___ relativistic _____________ (3) _ 90.0 _ -64.0 __ 41.5 ______ 5.97 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.79 _____ 6.01

_ 36 ___ Summer Blizzard ______ (4) _ 90.0 _ -64.0 __ 41.5 ______ 5.89 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.71 _____ 5.95

_ 37 ___ SLEETY _____________________ 88.0 _ -66.0 __ 42.8 ______ 5.65 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.47 _____ 5.65

_ 38 ___ s4lancia _______________ (2) _ 88.0 _ -66.0 __ 42.8 ______ 5.57 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.39 _____ 5.59

_ 39 ___ Moorlander ___________ (3) _ 88.0 _ -66.0 __ 42.8 ______ 5.49 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.31 _____ 5.53

_ 40 ___ Midlands Ice Age ___________ 87.0 _ -67.0 __ 43.5 ______ 5.29 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 5.11 _____ 5.29

_ 41 ___ 2010cold ___________________ 86.2 _ -67.8 __ 44.0 ______ 5.16 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.98 _____ 5.16

_ 42 ___ Mark Neal _(+ con) _________ 86.0 _ -68.0 __ 44.1 ______ 5.04 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.86 _____ 5.04

_ 43 ___ Stationary Front ____________ 85.0 _ -69.0 __ 44.7 ______ 4.92 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.74 _____ 4.92

_ 44 ___ Northeast Blizzard _____ (2) _ 85.0 _ -69.0 __ 44.7 ______ 4.84 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.66 _____ 4.86

_ 45 ___ Mulzy _______________________ 84.0 _ -70.0 __ 45.4 ______ 4.68 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.50 _____ 4.68

_ 46 ___ noname_weather ______ (2) _ 84.0 _ -70.0 __ 45.4 ______ 4.60 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.42 _____ 4.62

_ 47 __ Booferking ______________ (3) _ 84.0 _ -70.0 __ 45.4 ______ 4.52 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.34 _____ 4.56

_ 48 __ February 1978 _______________ 82.0 _ -72.0 __ 46.7 ______ 4.32 _ -0.10_ -0.02_ -0.06_ -0.00 ___ 4.14 _____ 4.32

_ 49 __ Dog Toffee __________________ 78.2 _ -75.8 __ 49.3 _______ 4.20 _ -0.10_ -0.03 _-0.06_ -0.00 ___ 4.01 _____ 4.20

_ 50 __ Polar Gael ___________________ 78.1 _ -75.9 __ 49.4 _______ 4.08 _ -0.10_ -0.03 _-0.06_ -0.00 ___ 3.89 _____ 4.08

_ 51 __ moffat _______________________ 78.0 _ -76.0 __ 49.5 ________3.95 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.76 _____ 3.95

_ 52 __ Emmett Garland _________ (2) _78.0 _ -76.0 __ 49.5 _______ 3.87 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.68 _____ 3.89

_ 53 __ pegg24 _______________________ 76.0 _ -78.0 __ 50.7 _______3.71 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.52 _____ 3.71

_ 54 __ cawsand _________________ (2) _ 76.0 _ -78.0 __ 50.7 _______3.63 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.44 _____ 3.65

_ 55 __ seaside60 ________________ (3) _76.0 _ -78.0 __ 50.7 _______ 3.55 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.36 _____ 3.59

_ 56 __ DR(S)NO ______________________ 75.0 _ -79.0 __ 51.4 _______3.35 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.16 _____ 3.35

_ 57 __ syed2878 ________________ (2) _ 75.0 _ -79.0 __ 51.4 _______3.27 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.08 _____ 3.29

_ 58 __ stewfox __________________ (3) _ 75.0 _ -79.0 __ 51.4 _______3.19 _-0.10_ -0.03 _-0.06_ -0.00 ___ 3.00 _____ 3.23

_ 59 __ J 10 ______________________ (4) _ 75.0 _ -79.0 __ 51.4 _______ 3.11 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.92 _____ 3.17

_ 60 __ stratty ________________________ 73.0 _ -81.0 __ 52.7 ________2.87 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.68 _____ 2.87

_ 61 __ Roger J Smith _________________72.7 _ -81.3 __ 52.9 _______ 2.74 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.55 _____ 2.74

_ 62 __ karlos1983 ___________________ 72.0 _ -82.0 __ 53.3 ______ 2.62 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.43 _____ 2.62

_ 63 __ Kirkcaldy Weather ____________70.4 _ -83.6 __ 54.4 ______ 2.50 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.31 _____ 2.50

_ 64 __ weather 26 _____________ (2) __ 70.4 _ -83.6 __ 54.4 ______ 2.42 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.23 _____ 2.44

_ 65 __ Walsall Wood Snow __________ 69.3 _ -84.7 __ 55.1 _______2.24 _-0.10_ -0.03 _-0.06_ -0.00 ___ 2.05 _____ 2.24

_ 66 __ Radiating Dendrite ___________ 69.0 _ -85.0 __ 55.3 _______2.12 _-0.10_ -0.03 _-0.06_ -0.00 ___ 1.93 _____ 2.12

_ 67 __ daniel* _______________________ 68.0 _ -86.0 __ 56.0 _______2.00 _-0.10_ -0.03 _-0.06_ -0.00 ___ 1.81 _____ 2.00

_ 68 __ Mr Maunder _________________ 66.0 _ -88.0 __ 57.3 _______1.88 _-0.10_ -0.03 _-0.06_ -0.00 ___ 1.69 _____ 1.88

_ 69 __ seabreeze86 ____________ (2) _ 66.0 _ -88.0 __ 57.3 _______1.80 _-0.10_ -0.03 _-0.06_ -0.00 ___ 1.61 _____ 1.82

_ 70 __ CheesepuffScott _____________ 63.0 _ -91.0 __ 59.3 _______1.64 _-0.10_ -0.04 _-0.06_ -0.00 ___ 1.44 _____ 1.64

_ 71 __ weather-history ______________62.0 _ -92.0 __ 60.0 _______1.51 _-0.10_ -0.04 _-0.06_ -0.00 ___ 1.31 _____ 1.51

_ 72 __ Let It Snow! __________________59.0 _ -95.0 __ 61.9 _______ 1.39 _-0.10_ -0.04 _-0.06_ -0.00 ___ 1.21*_____1.39

_ 73 __ Timmytour __________________ 56.0 _ -98.0 __ 63.9 _______ 1.27 _-0.10_ -0.04 _-0.06_ -0.00 ___ 1.10 _____ 1.27

_ 74 __ DAVID SNOW ___________ (2) _ 56.0 _ -98.0 __ 63.9 _______ 1.21 _-0.10_ -0.04 _-0.06_ -0.00 ___ 1.04 _____ 1.21

_ 75 __ sundog ______________________ 54.0_ -100.0 __ 65.2 _______ 1.03 _-0.10_ -0.04 _-0.06_ -0.00 ___ 0.88 _____ 1.03

_ 76 __ Steve Murr __________________ 52.0_ -102.0 __ 66.4 _______ 0.91 _-0.10_ -0.04 _-0.06_ -0.00 ___ 0.77 _____ 0.91

_ 77 __ prolongedSnowLover _______ 51.0_ -103.0 __ 67.1 _______ 0.79 _-0.10_ -0.04 _-0.06_ -0.00 ___ 0.66 _____ 0.79

_ 78 __ Leo97t ______________________ 50.0_-104.0 __ 67.7 _______ 0.67 _-0.10_ -0.04 _-0.06_ -0.00 ___ 0.55 _____ 0.67

_ 79 __ Norrance ___________________ 49.0_-105.0 __ 68.3 _______ 0.55 _-0.10_ -0.04 _-0.06_ -0.00 ___ 0.44 _____ 0.55

_ 80 __ Hotspur62 __________________46.0_-108.0 __ 70.2 _______ 0.43 _-0.10_ -0.05 _-0.06_ -0.00 ___ 0.33 _____ 0.43

_ 81 __ Neil N _______________________44.0_-110.0 __ 71.7 _______ 0.30 _-0.10_ -0.05 _-0.06_ -0.00 ___ 0.22 _____ 0.30

_ 82 __ DiagonalRedLine ______ (2) _ 44.0_-110.0 __ 71.7 _______ 0.22 _-0.10_ -0.05 _-0.06_ -0.00 ___ 0.14 _____ 0.24

_ 83 __ stargazer ___________________ 41.0_-111.0 __ 73.7 _______ 0.00 _-0.10_ -0.05 _-0.06_ -0.00 ___ 0.00 _____ 0.00

 

Note 72 _ (*) Scores from 72 to 83 descend in equal steps to zero with any duplicates noted. The full value of the deductions

is increasingly unused until last scores do not use any of the deductions to stay zero or next step(s) above. 

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

(confirmed results Jan 2)

The table was reorganized slightly on Dec 16 due to a few omissions of forecasters in the table of forecasts (page 5).

It was also recounted and the actual number of forecasts set at 83 plus the three robotic (normals, consensus) forecasts. These are scored in partial steps not affecting the scoring for actual participants. 

The table was edited for a higher outcome on 25th.

Edited December 29, 2020 by Roger J Smith

[Scorcher]

10 hours ago, Summer Sun said:

4.7c to the 12th

0.2c below the 61 to 90 average
0.4c below the 81 to 10 average

___________________________

Current high this month 5.1c on the 2nd
Current low this month 3.9c on the 8th & 9th

The only way is up now- mild week coming up should mean it's comfortably above average by the end of the week. Puzzled by some in this thread who seemed to think the outlook wasn't that mild? Double figures all the way in the south this coming week- could even come into the 'very mild' category.

[I remember Atlantic 252]

5 minutes ago, Scorcher said:

The only way is up now- mild week coming up should mean it's comfortably above average by the end of the week. Puzzled by some in this thread who seemed to think the outlook wasn't that mild? Double figures all the way in the south this coming week- could even come into the 'very mild' category.

Aye, second half seems to be warmer, often Atlantic more dominant from now until normally 1st week of March

[The PIT]

Sunny Sheffield up to 4.4C -1.6C, Rainfall 45.6mm 52.5% of the monthly average

[BornFromTheVoid]

Here's the projected values, with the GFS 7 day forecast included. The values below in brackets are from 2 days ago, to show the change in projections.

Warmest: 7.2C (7.4C)
Coldest:  2.3C (2.0C)
Top 10:  6.6C (6.5C)
Bottom 10:  3.4C (3.0C)
81-10:  5.3C (5.0C)

Quite a big change from 2 days back. There is now a 30% chance of finishing close to average (between 4.1C and 5.1C), 14% chance of below average (< 4.1C) and a 56% chance of above average (> 5.1C).
5.5 - 6.0C has the highest probability at the moment (24.2%).  A well above average month looking likely if we can't get that cold spell in the final week.

[Roger J Smith]

EWP projections are down slightly, in most part due to that cold, dry last week being advertised on latest GFS run. We are at about 65 mm after yesterday's addition to the 57 mm on the tracker to 12th. The ten-day GFS looks to be a solid 40-45 mm across the grid, with some western and northern parts well above that. So we should be close to if not above 110 mm by the 24th, but very little additional is shown so if all that came to pass, the provisional scoring would change for anyone with a forecast above 95 mm, some currently near the top would slide down and some others would move ahead into the top spots. 

If we consider the first ten days fairly reliable then it all depends on whether that cold, dry spell happens or not, which was the same conclusion being drawn above about the CET coming back down from what looks certain to be a gradual rise towards 6.0 at some point before Christmas. The current maps would argue for that coming down again towards 5.0 by end of month (an average of 1.6 for seven days at end of month turns a 6.0 into a 5.0). 

 

[The PIT]

Sunny Sheffield at 4.5C -1.4C below normal. Rainfall 45.8mm 52.7% of the monthly average

[Summer Sun]

5.1c to the 14th

0.1c above the 61 to 90 average
0.1c below the 81 to 10 average

___________________________

Current high this month 5.1c on the 2nd & 14th
Current low this month 3.9c on the 8th & 9th

[feb1991blizzard]

Still very confident with my 5.3 - obviously not confident of getting it bang on but surely wont be far away from the leaderboard - reckon i will be within 0.3 or 0.4.

[Roger J Smith]

I would think so also, the late cold is looking rather weak on the latest GFS run, would estimate 5.5 a very good number from current indications. That also tends to bring the EWP back up towards my earlier scoring table estimate (only needs to reach 124 mm for that to be generally valid, LG is a wild card appearing at some random location depending on exact outcome, to avoid dead last LG needs the outcome to exceed 120.5 mm (our lowest forecast was 41 mm), and he passes ten more low forecasts 44-56 mm by 128.1 mm. 

[The PIT]

Sunny Sheffield on 4.7C -1.1C below normal. Rainfall unchanged.

[Summer Sun]

5.2c to the 15th

0.3c above the 61 to 90 average
0.2c above the 81 to 10 average

___________________________

Current high this month 5.2c on the 15th
Current low this month 3.9c on the 8th & 9th

[Roger J Smith]

00z GFS run was cold and dry again after 23rd, 06z also rather cold but has a rebound for 30-31. These drier scenarios may limit EWP to 105-110 mm outcomes, we should easily get to 100 mm (currently around 75 mm) by 23rd but not guaranteed to add much after that. 

Would say most recent CET estimates are back down a bit to 5.0, highest running mean I foresee is 6.1 by Saturday, perhaps a plateau near that value for a week, then down steadily 23rd to 29th, checking at the end. 

[BornFromTheVoid]

Here's the projected values, with the GFS 7 day forecast included. The values below in brackets are from 3 days ago, to show the change in projections.

Warmest: 6.8C (7.2C)
Coldest:  3.3C (2.3C)
Top 10:  6.5C (6.6C)
Bottom 10:  4.0C (3.4C)
81-10:  5.5C (5.3C)

Big changes yet again. There is now a 32% chance of finishing close to average (between 4.1C and 5.1C), 6% chance of below average (< 4.1C) and a 62% chance of above average (> 5.1C).
5.5 - 6.0C has the highest probability at the moment , increasing from 24.2%  3 days ago to 31.9%.

[Roger J Smith]

EWP back on course now to surpass 125 mm, I have reworked the provisional scoring table (posted on Sunday 13th) to show the outcome for 130 mm.

At this point we are close to 80 mm, the tracker had reached 74 mm after 15 days and looks to have added about 6 mm yesterday (heavier rainfalls than that in the western third). The ten-day projection is a healthy 40 mm and could be closer to 50, then from 27th to 31st current (06z GFS) maps indicate potential for at least 15-20 mm more. The totals could therefore go as high as 150 mm. That would move LG steadily up through the bottom half of the scoring so your score (if not LG) could drop by 0.12 if you're between 40th and 70th place currently. LG could score anywhere from 1 to 5 in that range (130-150), as the error falls against the rising values of the rest of the field. Oddly everyone could gain marginally as the value goes up because the base for error adjustment will grow larger as the value of leader (SteveB)'s error increases, spreading out the zones of points adjustments. I guess that makes sense in this new tweaked scoring system (see post previous to that for a rundown) where the basic principle is that the leader scores 10 no matter what, then other scores are adjusted down from raw values increasingly for differentials, if the leader's error grows larger, then relatively other forecasts are probably going to have more similar errors in percentage terms. 

Anyway, the last few GFS runs have shown us probably four rather different outcomes beyond the similar looking first seven days (where the CET value is likely to peak around 6.0 at some point, yesterday's value was 5.4). Some runs have looked very cold for most of the period 23rd to 31st, others more up and down. The latest run is fairly cold in general there, but certainly not always so, the average during that period would probably stay above 3 C, a figure which gives an outcome of 5.1 C if the average on 22nd were to be 6.0. A value of 2.0 for 23-31 gives 4.9, can't see a northerly to northwesterly sourced cold spell going much lower than that, so 4.9 to 5.5 looks to be the most comfortable range at this point. It could go back up if that colder spell gets downgraded.

(for later use, if I need to update the CET tracker, 5.4 is 0.5 above 1961-90 mean, haven't looked at today's running 1981-2010 CET value yet, probably Summer Sun will update the tracker anyway. If not by midnight then I will update it. )

A January 2021 thread will open up on the 19th (Saturday). 

[Roger J Smith]

For continuity of the daily CET tracker ...

5.4c to the 16th

0.5c above the 61 to 90 average
0.4c above the 81 to 10 average

___________________________

Current high this month 5.4c on the 16th
Current low this month 3.9c on the 8th & 9th

[summer blizzard]

Is this the coolest first half since 2017.

[Roger J Smith]

Here are the CET values through 16th for past ten Decembers: 

2010 ___ 0.3 _ _ _ _ _ _ 2015 ___ 9.1

2011 ___ 5.0 _ _ _ _ _ _ 2016 ___ 7.0

2012 ___ 2.6 _ _ _ _ _ _ 2017 ___ 3.5

2013 ___ 6.7 _ _ _ _ _ _ 2018 ___ 6.6

2014 ___ 4.4 _ _ _ _ _ _ 2019 ___ 5.4

So we're sort of tied with last year except that our 5.4 is going to be a bit lower in the final numbers, possibly. (the 5.4 from last year is 5.35 rounded up).

In case you were surprised by 2010, that was the high point for the month, came after a few milder days mid-month. It was soon back down below -1.0.

And 2015 was close to the low point for that month, coming just before a run of record high daily (and in one case monthly) readings. 

[Scorcher]

3 hours ago, summer blizzard said:

Is this the coolest first half since 2017.

Not really something to remark about though, it's only 3 years ago and it's pretty much identical to last year!

[The PIT]

Sunny Sheffield at 5C -0.7C below average, Rainfall 54.9mm 63.2% of the monthly average.