Jump to content

  • Member Statistics

    17,609
    Total Members
    7,904
    Most Online
    NH8550
    Newest Member
    NH8550
    Joined

February 2013 Forecast Contest, Temperatures

Roger Smith

Recommended Posts

Roger Smith

Roger Smith

Posted
Posted

After a rather wild ride in January, it's time for your entries for February, and please note that we now have three new stations, Chicago (ORD) Atlanta (ATL) and Houston (IAH) in addition to our traditional three eastern locations, DCA, NYC and BOS.

 

The information in this post is mainly for new contestants as I expect most of the "regulars" are aware of the rules. For them, only this next part is probably worth a glance ... please submit your entries in the contest order and note for your convenience, any list of six forecasts across the page or down the page will be accepted without station identifiers if you want, and here's a template you could copy and convert to numbers:

 

DCA ... NYC ... BOS ... ... ORD ... ATL ... IAH

 

If you prefer to list them down the page, that's just as good (but slightly more work for me in making up a table of entries).

 

For new entrants (and one or two not so new) please note three things, contest order is important in avoiding scoring mistakes (thanks in advance) but I do make the effort to spot any scrambled order of forecasts. We are predicting anomalies (relative to 1981-2010 normals) and not actuals. Your entries should have one decimal and use the number zero for fractional anomalies (e.g., +0.2 rather than .2). Once again, I make the effort to decode every entry, but you'll see how much clearer the +0.2 version is. And third, I am recommending that people use a double minus that will look like this -- and not like that -. That's just because the font most people use here is rather small and I have "senior vision."

 

Late penalties

 

After a lenient start, I have proposed (without hearing any objections) to enforce the late penalty without exceptions in 2013, no matter what day of the week the first of the month falls, or holidays etc, and basically the only exception would be if (let's hope not) another event like Sandy delays entries due to emergency conditions. Then I would extend the deadline up to three days. Otherwise, I think it's fair to have a strict deadline policy, so that advantages gained by late entry are more than offset by penalties. And what is the penalty, you may ask? The same as last year, I think, one percent deduction per late hour per forecast made. Example, you're 17 hours late, then your six forecasts are scored off 83 rather than 100. There will be a five minute grace period for all penalties, timed by my forum time stamp (I have reset to eastern time just for this). That means you really have until 00:05 Feb 1st to enter without penalty and each hour or fraction after that counts (so your entry at 2:19 a.m. on the first is a 3% penalty, etc). Regardless of penalty and raw score, any score calculated to be below zero (negative) will be zero in the scoring table.

 

Scoring rules

 

The table of entries is created at the end of day four of each month (after which forecasts are not accepted mostly because they are already scoring zero anyway due to 96 hours of lateness). With the table of entries will be a notice of deadline for entrants to confirm their entries, changes will only be allowed on a good faith basis that the error is mine and not due to late editing. After that deadline (about the 7th) the table of forecasts is the official log of entries regardless of what may appear in your post (this prevents anyone from considering late editing and protects the field).

 

When actual values are in the range --5.0 to +5.0 scoring is simply a 2-pt deduction (from a base of 100 minus any late penalty) per 0.1 degree error.

 

When actual values are outside that range, an expanded scoring range comes into play, and it is summarized in this table (examples, other cases can be worked out from these guidelines):

 

ADJUSTED SCORING TABLE FOR LARGE ANOMALIES

_______________________________

 

examples are positive but same applies to large negative anomalies

 

Actual ............. A  (1-pt) ...... B zone (2 pt diff) ..... C zone (1-pt) ..... D zone (2-pt diff) ... E zone (1-pt)

___________________________________________________________________________________________

 

5.5 ...................0.1 .. 0.5 ......... 0.6 ... 5.0 .......... 5.1 .. .. 6.0 ....... 6.1 .. 10.5 ........... 10.6 ... 11.0

 

5.8 ..................0.1 .. 0.8 ..........0.9 ... 5.0 ............5.1 .. .. 6.6 ....... 6.7 .. 10.8 ............ 10.9 ... 11.6

 

6.0 ..................0.1 .. 1.0 ......... 1.1 ... 5.0 ........... 5.1 .. .. 7.0 ........ 7.1 .. 11.0 ........... 11.1 ... 12.0

 

6.5 ..................0.1 .. 1.5 ......... 1.6 .. 5.0 ........... 5.1 .. .. 8.0 ......... 8.1 .. 11.5 ........... 11.6 ... 13.0

 

7.0 ..................0.1 .. 2.0 .......... 2.1 .. 5.0 ........... 5.1 .. .. 9.0 ..........9.1 .. 12.0 ........... 12.1 ... 14.0

 

8.0 .................0.1 .. 3.0 ........... 3.1 .. 5.0 ........... 5.1 .. 11.0 .........11.1 .. 13.0 ........... 13.1 ... 16.0

 

9.0 .................0.1 .. 4.0 ........... 4.1 .. 5.0 ............ 5.1 .. 13.0 .........13.1 .. 14.0 ....... ... 14.1 ... 18.0

 

10.0 .............. 0.1 ... 20.0 (all 1-pt)

 

11.0 ............. 1.1 ... 21.0 (all 1-pt)

 

etc

 

Once again these scores are applied to a base of 100 minus penalty points.

 

----------

 

Once again, where scoring penalties are greater than 100 the score is not negative but zero.

 

This scoring method gives you the opportunity to check your scores against my calculations. Under the rules of official entries, you may request a scoring review at any time during the year ahead on any score you believe requires correction. The table of entries in the relevant thread remains the source of the official entry to be reviewed.

 

Contest results

 

I hope to score the monthly contests in a timely fashion and report the results within 2-3 days of end of each month, the NWS provisional CF6 reports are the source of our verification numbers and these are often reported on an ongoing basis in threads to keep you informed (most of you probably look at the source for that anyway). The contest winner is the person with the highest accumulated score for all six sites regardless of how many forecasts they submit. There will also be contest categories for the "classic" three stations of DCA/NYC/BOS and for the "expanded" trio of new stations we are predicting (ORD/ATL/IAH). If I get even more energetic I will announce winners for each of the six stations at the end of the year. The contest year has already started with the January contest.

 

Well, now that the rules are set out one time, we won't have to see them again so remember where they were posted (think Groundhog Day).

 

Good luck and have fun !!!

Link to comment
Share on other sites
  • Replies 72
  • Created
  • Last Reply

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...