Jump to content
  • Member Statistics

    15,497
    Total Members
    7,904
    Most Online
    easyaccess
    Newest Member
    easyaccess
    Joined
Sign in to follow this  
Roger Smith

2019 Hurricane (all named Tropical storms) forecast contest

Recommended Posts

What kind of North Atlantic tropical cyclone season do you predict for 2019? 

Our annual contest asks you to predict the total number of storms, and break it down by months. As you know, the "count" is total number of named storms, then total number of those that become hurricanes, and finally total number of hurricanes that become major (cat-3 or stronger). 

In 2017 the final count was 17/10/6 and our contest winner was Rockchalk83. In 2018 the final count was 15/8/2 and the contest winner was "A few Universes below normal" closely followed by UIWWildThing. Also in 2018, contest normal and the NHC mid-range forecast scored higher than any contest entrants. 

The rules are fairly simple. You need to post a seasonal total by the deadline which I am setting as June 1st but this could be extended a few days to increase the participation as long as no new named storms appear in early June. 

Your seasonal total should include any developments in May, and just as I thought of starting the contest, STS Andrea has appeared. The current forecasts for that cap it at 35 knots, so the count is already 1 0 0 for May. If there is further activity in May, that will be counted in the seasonal total too. But you don't have to predict anything for May, just add 1 0 0 for May in your forecast line. Your seasonal total will be adjusted if the monthly totals for your forecast when added up match your seasonal total but the 1 0 0 May portion changes. Contest normal will also change as that is based on the assumption that the 1989-2018 average used to create it applies to the defined season (this May activity is assumed to be extra to the normal values although May activity is becoming a fairly common occurrence recently).

In 2017, all entrants gave monthly predictions and almost all left them unedited, but the contest rules allow you to submit these up to 06z of the first of each month, or to revise those already submitted by that deadline. Note, there is no requirement for your monthly numbers to add up to your seasonal numbers and you can use decimal points to express uncertainty. In 2018, one entrant (the eventual contest winner) submitted only a seasonal forecast. When that happens, I take a scaled version of contest normal that adds up to the entrant's seasonal totals. Those were fairly similar in that case so the entrant was playing off the contest normal basically, reduced by one IIRC. 

I am going to post my forecast mostly as a guide to how your forecast should appear (not the numbers but the format) ...

Roger Smith ____ 18 12 4 ____ (May) 1 0 0 _ (Jun) 1 1 0 _ (Jul) 1 1 0 __ (Aug) 4 3 1 __ (Sep) 6 4 2 __ (Oct) 4 3 1 __ (Nov-Dec) 1 0 0

(this assumes a result of 1 0 0 May, would be adjusted to 19 12 4 if 2 0 0 or 19 13 4 if 2 1 0)

The following are contest normals (adjusted for the May 1 0 0 ) and the current (amended June 5th when one storm and one hurricane added to April forecast) CSU and NHC mid-range forecasts below them. Note that the monthly forecasts for these are scaled down versions of the contest normals, they are not actual forecasts, only seasonal totals were predicted. This is for contest scoring comparisons. I am also going to track NHC high end of range as I suspect that may do better than their mid-range. I am not totally sure if their forecasts include the May activity or are meant to apply to June to November only, so having the high end forecast in play will perhaps cover that ambiguity.

Contest Normal __16 8 3 ____ (May) 1 0 0 __ (Jun) 1 0 0 _ (Jul) 1 0 0 _ (Aug) 4 2 1 __ (Sep) 6 4 1 __ (Oct) 3 2 1 _ (Nov-Dec) 0 0 0

NHC (midrange) _ 12 6 3 ____ (May) 1 0 0 __ (Jun) 1 0 0 _ (Jul) 1 0 0 _ (Aug) 3 2 1 __ (Sep) 4 3 1 __ (Oct) 2 1 1 _ (Nov-Dec) 0 0 0

NHC (high end) __15 8 4 ____ (May) 1 0 0 __ (Jun) 1 0 0 _ (Jul) 1 0 0 _ (Aug) 4 2 1 __ (Sep) 5 4 2 __ (Oct) 3 2 1 _ (Nov-Dec) 0 0 0

CSU (Junel fcst) __13 6 2 ____ (May) 1 0 0 __ (Jun) 1 0 0 _ (Jul) 1 0 0 _ (Aug) 3 2 1 __ (Sep) 5 3 1 __ (Oct) 2 1 0 _ (Nov-Dec) 0 0 0 

Scoring for the contest is as follows:

50% of the score is based on the seasonal. You start with 50 points and in each category, you lose half of (1 point per error plus that number squared). Example, you predict 16 storms, the actual is 19, your error is (3 + 3 squared)/2 = 12/2 = 6. If you had similar errors for hurricanes and majors, your total score would then be 32/50. 

The other 50% of the score is based on your monthly forecasts starting with June and ending with Nov-Dec. These are worth 4, 6, 12, 16, 10 and 2 points in order from June to Nov-Dec. Then the errors reduce your possible score in the same way as the seasonal formula, except that June, July and Nov-Dec go with half the reductions (in other words, you lose points at half the rate of the other more active months). 

In the past few years, highest scores in the contest have been close to 90 (dropping to 84 last year) and seasonal scores of 48 to 50 have been achieved. A good monthly set has earned scores totalling 41-43 points. 

Good luck if you enter. The deadline will be made more precise around June 1st and all entrants can edit up to the eventual deadline without notice, you can assume that I won't be copying down or storing any forecasts until a firm deadline is posted in the thread and up to that firm deadline, so no need to post new numbers, just edit the old ones. Once a table of entries appears, your numbers are set for the seasonal and June portions but you have the option of posting revised monthly numbers at any point during the contest before monthly deadlines. Late monthly revisions will be penalized at a rate of 10% per day but will not be valid after any named storm is declared in the month. 

Share this post


Link to post
Share on other sites

16/7/3

May 1/0/0

June 1/0/0

July 3/0/0

August  3/2/1

September  6/4/1

October  2/1/1

Nov/Dec  0/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites

14/6/3

May 1/0/0

June 1/0/0

July 3/2/0

August 3/2/1

September 3/1/1

October 2/1/1

November/December 1/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites

17/7/3

May 1/0/0

June 2/1/0

July 2/1/0

August 3/1/1

September 5/3/2

October 4/1/0

November/December 0/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites

16/9/4

May 1/0/0

June 1/0/0

July 2/1/0

August 4/3/1

September 5/4/2

October 2/1/1

November/December 1/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites

13/5/2

May 1/0/0

June 1/0/0

July 2/1/0

August 3/1/0

September 3/2/1

October 2/1/1

November/December 1/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites

13/7/2

May 1/0/0

June 1/0/0

July 1/1/0

August 3/2/0

September 4/3/2

October 3/1/0

November/December 0/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites

The deadline for entries will be extended through the weekend of June 1-2 to encourage a larger field. Those who have already entered may edit current forecasts during this time.

Absolute deadline June 3rd 0600z, or earlier if the current slowly developing system depicted on NHS guidance map near Belize gets a more definite forecast for named storm activity.

Thanks for entering. 

Share this post


Link to post
Share on other sites

15/7/2

May 1/0/0

June 1/0/0

July 1/1/0

August 3/2/0

September 5/3/1

October 3/1/1

November/December 1/0/0

  • Thanks 1

Share this post


Link to post
Share on other sites
On 5/31/2019 at 7:18 PM, AfewUniversesBelowNormal said:

June 1

July 2

August 3

September 4

October 2

November 1

Okay so that's 13 more named storms, a total of 14 counting the one already recorded in May ... care to mention how many will be hurricanes and of those, major hurricanes? I will enter your numbers scaled to the contest normal in the absence of any further numbers. Thanks.

And in general, we have reached the deadline now, so a table of entries will be posted soon. 

Share this post


Link to post
Share on other sites

Table of entries for 2019 Atlantic Tropical-Season Forecast Contest

This year, will be keeping all the non-contest-entrant data in a separate zone at the top of the table. For entrants, the forecasts are listed in descending order of total storms with total hurricanes the second point of breaking ties, and majors the third point. 

FORECASTER ___________ SEASONAL __ MAY ___ JUN___ JUL ___ AUG ___ SEP ___ OCT ___ NOV-DEC

Contest Normal ___________16 8 3 _____ 1 0 0 __ 1 0 0 __ 1 0 0 __ 4 2 1 __ 6 4 1 __ 3 2 1 __ 0 0 0

Contest consensus ________ 15 7 3 _____ 1 0 0 __ 1 0 0 __ 2 1 0 __ 3 2 1 __ 5 3 1 __ 2 1 1 __ 1 0 0

NHC (high end) ___________15 8 4 _____1 0 0 __ 1 0 0 __ 1 0 0 __ 4 2 1 __ 5 4 2 __ 3 2 1 __ 0 0 0

NHC (midrange) __________ 12 6 3 _____1 0 0 __ 1 0 0 __ 1 0 0 __ 3 2 1 __ 4 3 1 __ 2 1 1 __ 0 0 0

CSU (June fcst) ___________13 6 2 _____1 0 0 __ 1 0 0 __ 1 0 0 __ 3 2 1 __ 5 3 1 __ 2 1 0 __ 0 0 0 

_____________________________________________________________________________________

Roger Smith _____________18 12_4 ____ 1 0 0 __ 1 1 0 __ 1 1 0 __ 4 3 1 __ 6 4 2 __ 4 3 1 __ 1 0 0

ineedsnow _______________17 7 3 _____ 1 0 0 __ 2 1 0 __ 2 1 0 __ 3 1 1 __ 5 3 2 __ 4 1 0 __ 0 0 0

RJay ____________________16 9 4 _____ 1 0 0 __ 1 0 0 __ 2 1 0 __ 4 3 1 __ 5 4 2 __ 2 1 1 __ 1 0 0

snowlover2 ______________ 16 7 3 _____ 1 0 0 __ 1 0 0 __ 3 0 0 __ 3 2 1 __ 6 4 1 __ 2 1 1 __ 0 0 0

NorthArlington101 _________16 7 3 _____ 1 0 0 __ 1 0 0 __ 3 1 0 __ 3 2 1 __ 5 3 1 __ 2 1 1 __ 1 0 0

Julian Colton _____________ 15 9 4 _____ 1 0 0 __ 0 0 0 __ 2 1 0 __ 2 2 0 __ 4 3 2 __ 4 2 2 __ 2 1 0

NCforecaster89 ___________ 15 7 3 _____ 1 0 0 __ 1 0 0 __ 2 1 0 __ 3 2 1 __ 4 3 1 __ 3 1 1 __ 1 0 0

hlcater __________________15 7 2 _____ 1 0 0 __ 1 0 0 __ 1 1 0 __ 3 2 0 __ 5 3 1 __ 3 1 1 __ 1 0 0

Stormlover74 ____________ 15 5 1 _____ 1 0 0 __ 0 0 0 __ 2 0 0 __ 3 2 0 __ 5 3 1 __ 3 0 0 __ 1 0 0

AfewUniv b n _____________14 9 3 _____ 1 0 0 __ 1 0 0 __ 2 1 0 __ 3 3 1 __ 4 3 2 __ 2 2 0 __ 1 0 0

Stebo ___________________14 6 3 _____ 1 0 0 __ 1 0 0 __ 3 2 0 __ 3 2 1 __ 3 1 1 __ 2 1 1 __ 1 0 0

CyclonicFury _____________ 14 6 2 _____ 1 0 0 __ 0 0 0 __ 1 0 0 __ 4 2 1 __ 5 3 1 __ 2 1 0 __ 1 0 0

yoda ____________________13 7 2 _____ 1 0 0 __ 1 0 0 __ 1 1 0 __ 3 2 0 __ 4 3 2 __ 3 1 0 __ 0 0 0

yotaman ________________ 13 5 2 _____ 1 0 0 __ 1 0 0 __ 2 1 0 __ 3 1 0 __ 3 2 1 __ 2 1 1 __ 1 0 0

__________________________________________________________________________________

Apart from my somewhat higher forecast, the rest of our 14 entrants have chosen values fairly close to the long-term average with a bit of a spread, and so the contest is likely to be closely fought (unless we get a very active season). Our consensus is very close to the top of the NHC range. 

 

 

 

Share this post


Link to post
Share on other sites

June had no activity so the scoring (out of four possible points) looks like this:

4 pts (0,0,0 forecasts) __ Julian Colton, Stormlover74, Cyclonic Fury

3.5 pts (1,0,0 forecasts) _ all others except below

3.0 pts (1,1,0 forecast) __ Roger Smith

2.5 pts (2,1,0 forecast) __ ineedsnow

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

July may get some action apparently. 

Share this post


Link to post
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
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×