Jump to content

  • Member Statistics

    17,507
    Total Members
    7,904
    Most Online
    SnowHabit
    Newest Member
    SnowHabit
    Joined

2014 NATL Tropical-Hurricane-Seasonal Forecast Contest

Roger Smith

Recommended Posts

  • Replies 193
  • Created
  • Last Reply
Roger Smith

Roger Smith

Posted
  • Author
Posted

3/1/1.

Note: I wouldn't be surprised if my only forecasted "major" hurricane doesn't develop until the month of September. If 93L unexpectedly achieves hurricane intensity, this monthly forecast will bust.

 

Under contest rules, any storm that gets a name in late August will count towards your monthly forecast even if the hurricane or major hurricane stages don't begin until September. With that in mind, I tweaked the monthly normals thinking that August might gain, while September might stay about as per NHC climatology losing to August and gaining from October, so that October might be the month that would lose a bit under the contest method, as compared to the normals. However, I am still somewhat confused by how the NHC counts storms in months, it seems either (1) they assign all phases to the month when a storm gets numbered, or (2) they assign each phase to the month when it first develops. Since I can't find much about this, I have calculated the normals from a study of older climatology adjusted to the 1989-2013 seasonal normals. In any case the "normal" values entered into the contest are smoothed and not actual numbers from those seasons, but they add up to the seasonal normals.

 

Okay, I see that it is now 01z August 1st, no designation yet of TD 3 let alone a named storm, so this current candidate is definitely an August contest player for certain.

Link to comment
Share on other sites
Roger Smith

Roger Smith

Posted
  • Author
Posted

True, Bertha counts for August. However, let's say it went into July and only achieves TS status. The difference in scores for most of us would be 0.5 because July is being scored on half-scale error points.

 

In fact, the table of entries and updated scoring is now available for you to check, back at post 50 (June 3rd) as I have edited that table up to the entry deadline of 06z. Any later forecasts will be edited in, right now you'll find a provisional forecast based on a pro-rated portion of seasonal "normals" which in themselves are contest-adjusted -- August climatology may be more like 3/1/0 rounded off than the 4/2/1 that I have assigned, it's probably close to 3.5/1.5/0.5 but with our rules draining some of September's action, I've gone with 4/2/1 as the monthly "normal" for the contest, then whatever your seasonal is, relative to the 25-year average, determines what your provisional forecast is. Most of them are in the 3/1/1 range, some even lower (2/1/0 is about the lowest I calculated). All are of course subject to being changed or confirmed as they stand by yourself, if you post and become active for the month.  That's to your advantage even with the penalty clock now ticking (1% an hour late) as provisionals are given 50% of points scored. And a reminder, in August, we go to full error points. That means a two-storm error in any category removes (1+4)/2 or 2.5 points from the 12 available. In June and July that reduction was 1.2 (and 0.5 for a one-storm error, now it will be the full point).

 

Link to comment
Share on other sites
GaWx

GaWx

Posted
Posted

True, Bertha counts for August. However, let's say it went into July and only achieves TS status. The difference in scores for most of us would be 0.5 because July is being scored on half-scale error points.

 

In fact, the table of entries and updated scoring is now available for you to check, back at post 50 (June 3rd) as I have edited that table up to the entry deadline of 06z. Any later forecasts will be edited in, right now you'll find a provisional forecast based on a pro-rated portion of seasonal "normals" which in themselves are contest-adjusted -- August climatology may be more like 3/1/0 rounded off than the 4/2/1 that I have assigned, it's probably close to 3.5/1.5/0.5 but with our rules draining some of September's action, I've gone with 4/2/1 as the monthly "normal" for the contest, then whatever your seasonal is, relative to the 25-year average, determines what your provisional forecast is. Most of them are in the 3/1/1 range, some even lower (2/1/0 is about the lowest I calculated). All are of course subject to being changed or confirmed as they stand by yourself, if you post and become active for the month.  That's to your advantage even with the penalty clock now ticking (1% an hour late) as provisionals are given 50% of points scored. And a reminder, in August, we go to full error points. That means a two-storm error in any category removes (1+4)/2 or 2.5 points from the 12 available. In June and July that reduction was 1.2 (and 0.5 for a one-storm error, now it will be the full point).

 

 Roger,

 We still have til 2 AM EDT/6Z to avoid penalty. I'll be getting my August guess out shortly.

Link to comment
Share on other sites
Roger Smith

Roger Smith

Posted
  • Author
Posted

No problem, in fact just checking the rules, the late penalties are 10% a day for monthly forecasts just to encourage people to enter a bit late without taking an enormous hit. So the way that would work is a 1% score reduction every 2.4 hours into August plus 6h in z time.

 

But to make it more user-friendly, we'll just say no penalty for anyone who shows up by 10z then 1% every 2h after that.

 

Now is that going to change scores much? Let's say you got 9/12 for your August forecast and had a 6% penalty. Your score would be 8.5 instead of 9.0 ... so basically you're always further ahead to post a late forecast than to take the provisional 50% score.

 

If you post after this, check the table the following day to make sure I edited you in. If you edit a post already incorporated into the table, bring it to my attention and we'll see if there's a late penalty to be factored in, or not.

Link to comment
Share on other sites
  • 4 weeks later...
Roger Smith

Roger Smith

Posted
  • Author
Posted

The deadline for posting a September monthly forecast will be extended due to the Labor Day weekend to 06z Tuesday 2nd. After that, 1% per hour reductions in score will apply to forecasts entered.

 

As always, please note that storms named before 00z September 1st will count for August and all of their subsequent exploits in September will be scored relative to August forecasts. Storms that form in late September will similarly have all their history scored against your September forecast.

 

This is the most valuable month for scoring, worth 16 of the 50 available points. Anyone with a seasonal forecast who does not enter a monthly gets a provisional scaled against the normal values (which are contest adjusted) and that is scored at 50%. So you would gain points by doing that calculation yourself if you don't have any other forecast ideas.

 

I will enter with 4/3/2 but may edit that before the deadline. If you edit yours before the deadline, no need to make a note, I only start updating the tables after the deadlines. If you decide to edit after the deadline despite the penalty, once again, I do make a second tour of the entries around the 10th to check my work, but you'd be wise to note the edit.

 

Scoring will be updated (the post with entries and scoring is back around June 4 in this thread) when Cristobal and any further action have run their course. At this point August stands at 2/1/0 2/2/0. If that mid-ocean yellow X (or anything else) reaches named status before Sunday evening, its entire life cycle will be awaited for that scoring update.

 

Good luck with your September forecasts and as far as updating seasonals, if you want to make any comments go ahead but for contest purposes there are probably no scores except possibly Normal 1995-2006 if that was an entry, that could be mathematically improved under the rules (1% reduction in score per day) at this point, so updates will not be entered into the scoring table.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

Go to topic listing

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...