Advertisement
If you have a new account but can't post, please email Niamh on [email protected] for help to verify your email address. Thanks :)
New AMA with a US police officer (he's back!). You can ask your questions here

January 2018 Boards weather forecast contest

  • 28-12-2017 2:06am
    #1
    Registered Users Posts: 13,185 ✭✭✭✭ M.T. Cranium


    Welcome to the 2018 Boards weather forecast series ... we hope to see a few new entrants this year as well as all the regulars back for more of the MTC ritual torture (much of it self-inflicted as it turns out ...) :D:eek:

    So if you're a regular you won't read this and if you're new you may need to read it twice ...

    We ask five routine questions and add a bonus, and in this instalment I am adding the points structure for your interest.

    1. Predict the IMT ("Irish Mean Temperature") which is the average for five selected locations (1981-2010 averages in brackets), namely Claremorris (4.9 C), Mullingar (4.6 C), Casement and Carlow Oak Park (both 5.1) and Shannon (6.0) for an overall "normal" mean of 5.1 ... this part is worth 25 points and you lose one point per 0.1 deg error.

    2. Predict the MAX (highest temperature from recorded daily highs from the 25 regular reporting stations on the met.ie website), the normal value is set at 13.0. This is worth 20 points and errors of 0.1 reduce your score from that value.

    3. Predict the MIN (lowest temperature ... see 2 above ... value is set at -7.0 and this is also worth 20 points, same error reductions.

    4. Predict the PRC (precip in mm, which includes the liquid equivalent of snowfall on average 10 cm snow reduces to 10 mm precip, and we use eleven stations to determine this average, the five mentioned in part 1 above, and also Malin Head, Belmullet, Ballyhaise, Cork, Valentia, and Johnstown Castle. Average amounts for your guidance would be 60 to 150 mm, probably around 110 mm for the eleven station mean, but in any case your forecast should be a percentage not an actual amount. 100% would mean you expect normal amounts, 200% means twice normal etc. This part is worth 15 points and you lose 1 point for every 3% (rounded off) of error, so let's say you predict 90% and reality is 106%, that's 16/3 or 5 (rounded off) deducted so you score 10.

    5. Predict the SUN which is also a percentage of normal sunshine at six selected locations (Belmullet, Casement, Dublin, Cork, Shannon, Valentia). In January, the daily average is near 2 hours a day to give you some idea, there doesn't have to be a summery spell to push above that. This portion is worth 10 points, and errors are calculated in increments of 5% rounded off; for example, if you predict 80% and reality is 92%, then you lose 12/5 points which rounds off to 2, and you score 8.

    6. Each month there is a bonus question worth 10 points. These are scored in rank order and the absolute error does not matter much, although I use it to determine how to cluster the scores when I need to have more than an equal number on one score.

    This month (the adept regulars will have scrolled down to here to avoid my lecture above) the bonus question asks you to predict the strongest sustained wind and gust speed in knots. Locations for both (they could be different) will be worth a point each if correct and the two values will be scored out of 4. You can earn superbonus points here -- 4+4+1+1 will get 2 superbonus, any combination that reaches nine will get 1 superbonus.

    Here's my forecast (which I may edit by the deadline), use this as a template but change all the numbers or locations to those you want to predict.

    It helps me organize the tables faster if people use this template as shown (name first) ...

    M.T. Cranium _________ 4.2 __ 13.3 __ --8.0 __ 115 __ 119 __ 55,80 (Mace, Malin)

    (that last part means I expect a sustained wind max of 55 knots at Mace Head and a top gust to 80 knots at Malin Head.)

    Note: The wind speeds must occur at one of the 25 regular met.ie sites, not Fastnet Lighthouse or your hand-held aneomometer on top of a mountain. :)


    Deadline for on-time entries is always 0300h of 1st day of each month.

    Late penalties are going to be slightly relaxed for 2018. They were running 1 point per 3h late, now it will be 1 point per 4h late to 1500h 2nd followed by 1 point per hour after that (9 points would be accumulated by then).

    And for New Years, I will further relax the late penalties to 1 point any time before 1500h, then further points every six hours , so it's only going to be five accumulated points by 2nd (1500h), however, any prediction of MAX, MIN or wind speed that matches a value available to a late entrant will attract the same penalties. This continues on with the larger penalties after the 3rd so if you wanted to wait until let's say the 6th to nail those wind predictions, you would be looking at such massive penalties that you might score negative, and when people do that, we have a minimum score of 10/100 for any entry.


    Regulars also know these details -- there are two annual competitions, one is the Boards award which is based on your ten best scores out of however many you have (10, 11 or 12 are needed to qualify). Then there's the MTC Award going to the highest total score, you can try your luck with 11/12 but it usually takes the full set to top the field in this one. And we keep track of seasonal efforts too, unfortunately for new talent, December 2017 counts towards winter in this four seasons award and you're already d.q.'d from that season if you didn't enter December.

    One final note (again known to our regular crew) ... scores for any element that fall below a defined "minimum progression" get boosted to that minimum progression. This is to equalize monthly scoring for the purposes of dropping two low scores, we don't want everyone to be dropping the same months. However, you can gain an advantage here by scoring within 2 points of the maximum in any category, then if others need a boost, they only get half of the usual increase. Here's an example, let's say rainfall is only 40% and almost everyone predicts above 80, but one clever soul predicts 46%. They get a raw score of 13/15. With minimum progression, that's boosted to 15, but every other boost is cut in half. The next closest forecast might be 81% which would be worth only 1 point in raw scoring. Normally that would go to 14, but in this case the 13 point boost is cut in half to 7 and they get 8 points. Etc etc. This has happened once or twice, it rewards those who are daring and get it close. But in my example, if somebody had gone for 56% (raw score 10/15) then while they would still get the max (15) the next closest would go to 14. (note also, for SUN the highest score must be 9/10 or 10/10 to activate the reduced min prog; for bonus, this feature does not apply; also, minimum progression cannot lower your raw score so it can be applied to some portions of the scoring).

    Well, good luck and I hope to see some new entrants enjoying our fun contests, you'll be surprised at the skill level of Con Sensus, gender unknown, often quite close to the mark, NormaL not so much (except maybe this month of Dec 2017).

    Scoring is posted within 3-4 days of month end and is based on official data provided in Met Eireann's "Monthly Summary" as well as notes taken from daily and weekly summaries.


«1

Comments

  • Registered Users Posts: 622 ✭✭✭ Tae laidir


    Stations on the East Coast are generally drier than those on the West.
    In January, Baldonnel expects about 65mm and Valentia 175mm.
    Were we to have an extra 65mm at Baldonnel, would that increase Baldonnel's contribution to 200%?
    However, if that same 65mm were to fall at Valentia, would that increase Valentia's total to just 137% (175+65)*100/175)?
    Or do you just add all the rainfall totals together for the entire country?

    In the first half of the year, there is less daylight in the first half of the month. In March this is about 2 hours in each station between the 1st & 30th March.
    Perhaps, I were to think that we were to get unending sunshine from 1-15 March, and totally cloudy from 16-30 March.
    Would I propose that there would be exactly 50% sunshine, or would I expect less than 50%, as there is greater daylight in the second half of the month?

    And how many Angels can dance on the head of a needle?


  • Registered Users Posts: 1,713 ✭✭✭ Bsal


    Bsal _________ 4.9 __ 12.8 __ --7.0 __ 125 __ 102 __ 53,71 (Belmullet, Mace Head)


  • Registered Users Posts: 350 ✭✭ waterways


    Tae laidir wrote: »
    And how many Angels can dance on the head of a needle?

    To answer your last question:
    plenty, uncountable many when holding hands :)


  • Registered Users Posts: 4,995 ✭✭✭ John mac


    john mac_________ 4.9 __ 12.9__ -5.9 __ 105__ 98 __ 46,70 (Mace, knock)


  • Registered Users Posts: 12,554 ✭✭✭✭ sryanbruen


    sryanbruen _________ 4.6 __ 13.2 __ -8.0 __ 110 __ 90 __ 50,70 (Mace, Malin)

    Weather and climate site - https://www.ukandirelandclimate.com/ (advised to view on PC, not optimised for mobile)

    Photography site - https://www.sryanbruenphoto.com/



  • Advertisement
  • Registered Users Posts: 317 ✭✭ kindredspirit


    Kindred Spirit. ___ 4.9 ___ 12.9 ___ -7.0 ___ 112 ___ 92 ___ 50, 80 (Malin Head, Belmullet.)


  • Registered Users Posts: 457 ✭✭ dasa29


    dasa29 _________ 5.0 __ 13.5 __ --6.0 __ 120 __ 100 __ 45,55 (Sherkin island, Malin Head)


  • Registered Users Posts: 1,745 ✭✭✭ Joe Public


    Joe Public _________ 2.9 __ 13.6 __ -9.4 __ 102 __ 115 __ 57,78 (Belmullet, Malin)


  • Registered Users Posts: 164 ✭✭ sunflower3


    sunflower3 _________ 4.8 __ 13.4 __ -6.0 __ 120 __ 95 __ 55, 68 (Malin, Malin)


  • Moderators, Home & Garden Moderators, Science, Health & Environment Moderators Posts: 16,102 Mod ✭✭✭✭ DOCARCH


    DOCARCH _________ 4.3 __ 14.1 __ -7.1 __ 85 __ 110 __ 52,69 (Malin, Malin)


  • Advertisement
  • Registered Users Posts: 9,183 ✭✭✭ sdanseo


    sdanseo _________ 4.0 __ 13.3 __ -9.8 __ 103 __ 111 __ 53,70 (Mace, Belmullet)


  • Registered Users Posts: 359 ✭✭ mrskinner


    Skinner____5.3____13.7____-5.8___115___115___49,68(Malin,Malin)


  • Registered Users Posts: 19,554 ✭✭✭✭ Rikand


    rikand ..... 5.0 ..... 13.5 ..... -6.0 ..... 90 ..... 120 ..... 58, 85 (mace, sherkin)


  • Closed Accounts Posts: 6,466 Lumi


    Lumi _________ 4.5 __ 14.1 __ -8.5 __ 110 __ 90 __ 51,79 (Malin Head, Sherkin Island)


  • Registered Users Posts: 5,847 ✭✭✭ pauldry


    Pauldry_________ 5.2__ 14.3 __ -6.5__ 108 __ 101 __ 50,75 (Malin Head, Mace Hd)


  • Moderators, Science, Health & Environment Moderators Posts: 4,406 Mod ✭✭✭✭ mickger844posts


    mickger844posts _________ 4.8 __ 13.5 __ --7.0 __ 110 __ 115 __ 56,72 (Mace, Mace)


  • Registered Users Posts: 1,408 ✭✭✭ Jpmarn


    Jpmarn _________ 4.3 __ 13.5 __ -6.5 __ 125 __ 90 __ 58,78 (Belmullet, Dublin Airport)


  • Registered Users Posts: 13,185 ✭✭✭✭ M.T. Cranium


    I think it's better to take an average of station averages of precip or sunshine, rather than adding them all up and comparing the sum to a normal sum.

    The fact that one location normally gets more rain or sunshine than other locations does not affect the concept that you're reporting on the mean performance of these locations relative to normal.

    There are cases where you would not want to do this in fields other than climate. For example, if you wanted to calculate the percentage of Irish people who were natural redheads, you would be better to take the total number of redheads vs the total population, and not some list of county proportions.

    But that is due to the fact that there might be a million people in one county and just fifty thousand in another. With rainfall or sunshine, there are not more locations at the sunny or wet spots than elsewhere, it's just one location. And the grid is selected to provide a fairly representative sample of the country.

    Here's an example of what could happen if we went by percentages of totals and not the average of averages. Let's say the grid of eleven stations had one location with twice the average rainfall of all the others, and those ten were the same amount.

    Then in one particular month, it just doesn't rain at all at the wet place, and there is 110% at all the others. With average of average, that comes out to 100%, while in the total method, you come to just 91%. Now both of those are probably too low to be considered truly representative, after all 89% of locations experienced 110% of average rainfall. But the 100% is not as far off as the 89%. However, if your aim was to assess the total runoff into the ocean then perhaps the absolute total method would make more sense.

    (my forecast is back in post one, think I went cool, a bit dry and a bit sunny)

    Happy New Year to you all, I have to wait a few more hours here. Sun just setting behind the mountains.


  • Registered Users Posts: 622 ✭✭✭ Tae laidir


    Tae laidir_________ 4.9 __ 14.2 __ -7.3 __ 130 __ 94 __ 53,82 (Malin, Mace)

    Thanks MT.
    On average, averaging averages is above average!

    But a sunnier end fortnight in Spring is of greater value than the first fortnight?
    And a sunnier first fortnight in Autumn is of greater value than the end fortnight?


  • Registered Users Posts: 1,772 ✭✭✭ dacogawa


    Dacogawa ____ 4.7 ____ 13.7 ____ -6.9 ____105 ____97 ..... 52, 69 (Malin, Malin)


  • Advertisement
  • Registered Users Posts: 1,772 ✭✭✭ dacogawa


    Dacogawa ____ 4.7 ____ 13.7 ____ -6.9 ____105 ____97 ..... 52, 69 (Malin, Malin)


  • Registered Users Posts: 350 ✭✭ waterways


    Morning, I'm still counting angels and forgot my numbers :angel:
    waterways _________ 4.7 __ 13.7__ --7.1 __ 134 __ 100 __ 51,75 (Mace, Belmullet)


  • Registered Users Posts: 1,894 ✭✭✭ BLIZZARD7


    BLIZZARD7 ____ 1.7 ____ 14.4 ____ -13.2c ____90 ____120___ 60, 86 (Mace, Mace)


  • Registered Users Posts: 3,756 ✭✭✭ rameire


    Rameire. 4.8. 13.4. -4.9. 115. 89. 55, 72 Malin, Valencia

    Firstly apologies for the spacing, am on an irregular machine.
    Secondly, apologies for the late posting, am currently in NYC in constant minus temps and it’s playing havoc with the mind
    Thirdly. Merry New Year.
    Fourthly. Thank you M.T. Cranium for your consistent / persistent hard work.


  • Registered Users Posts: 1,745 ✭✭✭ Joe Public


    For the bonus question, I notice Mace Head is off to a good start on the Met eireann 5pm reports today

    MACE HEAD GALWAY(A) Wind Dir W Speed(Kts) 53 Gust 75


  • Moderators, Home & Garden Moderators, Science, Health & Environment Moderators Posts: 16,102 Mod ✭✭✭✭ DOCARCH


    Joe Public wrote: »
    For the bonus question, I notice Mace Head is off to a good start on the Met eireann 5pm reports today

    MACE HEAD GALWAY(A) Wind Dir W Speed(Kts) 53 Gust 75

    And Newport too at 1800, gust of 75kn.


  • Moderators, Home & Garden Moderators, Science, Health & Environment Moderators Posts: 16,102 Mod ✭✭✭✭ DOCARCH


    Gust of 84kn at Knock AP at 1900. :eek:


  • Registered Users Posts: 17,664 ✭✭✭✭ JCX BXC


    Forgot, ah well. At least MT has relaxed the late penalties.

    JCXBXC_________ 4.5 __ 13.6 __ -9.0 __ 110 __ 90 __ 50,84 (Mace, Knock)


  • Registered Users Posts: 2,569 ✭✭✭ 200motels


    No internet so that's my excuse.

    200motels
    4.7 13. -6.9. 100. 84. 54, 70 Malin, Valencia


  • Advertisement
  • Registered Users Posts: 13,185 ✭✭✭✭ M.T. Cranium


    January 2018 -- Table of Forecasts


    FORECASTER ________ IMT _MAX _MIN _PRC _SUN ___ Max winds and gusts

    MrSkinner ___________ 5.3 _13.7 _--5.8 _115 _115 ___ 49 Malin ___68 Malin
    Pauldry______________5.2 _14.3 _--6.5 _108 _101 ___ 50 Malin ___75 Mace

    NormaL _____________ 5.1 _13.5 _--7.0 _100 _100 ___50 (*) ____ 80 (**)

    rikand ______________ 5.0 _13.5 _--6.0 _090 _120 ___ 58 Mace ___85 Sherk
    dasa29 _____________ 5.0 _13.5 _--6.0 _120 _100 ___ 45 Sherk __ 55 Malin
    Tae laidir ____________4.9 _14.2 _--7.3 _130 _094 ___ 53 Malin ___82 Mace
    john mac ____________4.9 _12.9 _--5.9 _105 _098 ___ 46 Mace ___70 Knock
    Kindred Spirit. ________4.9 _12.9 _--7.0 _112 _092 ___ 50 Malin ___80 Bel
    Bsal ________________4.9 _12.8 _--7.0 _125 _102 ___ 53 Bel ____ 71 Mace
    mickger844posts ______4.8 _13.5 _--7.0 _110 _115 ___ 56 Mace ___72 Mace
    sunflower3 __________ 4.8 _13.4 _--6.0 _120 _095 ___ 55 Malin ___68 Malin
    Rameire _____ (-3) ___ 4.8 _13.4 _--4.9 _115 _089 ___ 55 Malin ___72 Val

    Con Sensus __________4.7 _13.5 _--7.0 _110 _100 ___ 52 Malin ___ 75 Malin

    Dacogawa ____ (-1) ___4.7 _13.7 _--6.9 _105 _097 ___ 52 Malin ___ 69 Malin
    waterways ____ (-1) __ 4.7 _13.7 _--7.1 _134 _100 ___ 51 Mace ___ 75 Bel
    200motels ___ (-25) ___4.7 _13.0 _--6.9 _100 _084 ___ 54 Malin ___70 Val
    sryanbruen __________ 4.6 _13.2 _--8.0 _110 _090 ___ 50 Mace ___70 Malin
    Lumi _______________ 4.5 _14.1 _--8.5 _110 _090 ___ 51 Malin ___79 Sherk
    JCXBXC ____ (-20) ___ 4.5 _13.6 _--9.0 _110 _090 ___ 50 Mace ___ 84 Knock
    DOCARCH ___________ 4.3 _14.1 _--7.1 _085 _110 ___ 52 Malin ___69 Malin
    Jpmarn _____________ 4.3 _13.5 _--6.5 _125 _090 ___ 58 Bel ____ 78 DubA
    M.T. Cranium ________ 4.2 _13.3 _--8.0 _115 _119 ___ 55 Mace ___80 Malin
    sdanseo ____________ 4.0 _13.3 _--9.8 _103 _111 ___ 53 Mace ___70 Bel
    Joe Public ___________ 2.9 _13.6 _--9.4 _102 _115 ___ 57 Bel ____ 78 Malin
    BLIZZARD7 ___ (-2) ___1.7 _14.4 _-13.2 _090 _120 ___ 60 Mace __ 86 Mace

    ____________________________________________________________

    23 entries so that the median (Con Sensus) is the 12th ranked value.

    Late penalties were relaxed at least for the first 36h, after that they would have amounted to even more than I assessed two players who entered on the 3rd, but I think these will be fair given the breaks given the earlier late forecasts. However, since JCX BXC had access to data from the 2nd, his bonus score is capped at 8 with no superbonus ... 200motels has no restrictions since he was either unaware of the marks set on the 2nd or is playing fair with lower values.

    For NormaL, the * and ** indicate that scoring for location will be based on whether the outcome is normal or less so, I will fit the points to what usually happens. For Con Sensus, here are the counts for max wind forecasts, with Con's points if these locations verify. If none of them verify, as with your forecasts, there will be partial scoring if you name 2nd or 3rd highest location (over the month).

    Station ___ Max wind ___ points ___ Max gust ___ points

    Malin ________10 ______ 2.5 ________ 8 ________ 2.5
    Mace ________ 9 _______2.0 ________ 5 ________ 2.0
    Belmullet _____ 3 ______ 1.5 ________ 3 ________ 1.5
    Sherkin Island _ 1 ______ 1.0 ________ 2 ________ 1.0

    any other _____ 0 ______ 0.0 ________ 4 ________ 0.5 or 1.0

    The other locations chosen (only for gusts) were Knock, Valentia and Dublin. Knock had 2 but one was disqualified for this purpose. Valentia had 2 votes and will score 1.0 if it verifies (for Con). Knock or Dublin will score 0.5.

    Our consensus forecast is for a rather cold January with somewhat above average precipitation (hopefully including snow?) and near average sunshine.

    Values of 53 (Mace) and 84 (Knock) were established on 2nd for the bonus question. In general, this did not influence the forecasts very much if at all, but either of these could of course be surpassed anyway.

    Good luck -- the 2017 final scoring has been posted.


Advertisement