Jump to content

  • Member Statistics

    17,606
    Total Members
    7,904
    Most Online
    ArlyDude
    Newest Member
    ArlyDude
    Joined

§ June Forecast Contest, Tropics §

Mallow

Recommended Posts

Mallow

Mallow

Posted
Posted

Rules are in the following thread:

Normally, I am going to make forecasts due by the end of the day on the second-to-last day of the month, but since this month was started late, forecasts are due by 11:59 PM EDT, Wednesday, June 1st.

I need a forecast for June's storms, canes, and majors, as well as your guess for the seasonal total storms, canes, and majors. The breakdown of points available this month (out of a total potential of 1000 for each variable for all seven months) are as follows...

POINTS AVAILABLE FOR JUNE

MONTHLY

Storms: 92

Canes: 45

Majors: 45

SEASONAL

Storms: 271

Canes: 261

Majors: 267

Remember, you can format your forecast to the nearest tenth (so you can, for example, forecast 5.3 hurricanes). I encourage you to copy and paste the format below to make your forecast.

JUNE

Monthly: S/C/M

Seasonal: S/C/M

Averages

Link to comment
Share on other sites
  • Replies 68
  • Created
  • Last Reply
Mallow

Mallow

Posted
  • Author
Posted

Mallow..., Im confused on how some forecasts have decimals. How can you get 5.3 hurricanes?

i suppose its for tiebreakers. might be better to use number of hurricane days for that.

june 1/0/0

season 13/8/4

No.

It's all in the rules.

For strategic reasons, you can give a forecast to the nearest tenth. Obviously, verification will be in whole numbers, but tenths can still be statistically advantageous in cases where you're split between two possible whole-number guesses.
Link to comment
Share on other sites
Inudaw

Inudaw

Posted
Posted

I've never had a problem with decimal forecast to play the odds for the overall forecast at the end verses trying to score a coop with the exact forecast. XD However this goes back the two semester worth of forecasting contest in my forecasting classes during my time at Penn State. We used a probabilistic based forecast system though in which we put down our confidence in categorical potential values verses laying down a deterministic forecast...

Example in how it would relate to tropical forecast.. (It was graded using a Root Mean squared deviation summation formula... where the lower the the sum of square deviations then the higher the score was ) Threfore a perfect forecast (but risky one.. ) would yield a deviation of 0. Where you forecast a 100 percent (1 :P ) chance of this number happening and it actually verifies. Conversely if you get it wrong..and you did not forecast any probably for the category that actually verified. (say 2 storms form and june... and you said that there was no chance for 2 storms to form in june. Your forcast of 0 percent chance for 2 storms to form would yield a value of 1 for that particular form.

(1-0)^2 = 1 :P You want this value to be 0 XD

more relastic case... For june say this is your probabilistic forecast

0 0.1 0

1 0.8 1

2 0.1 0

(0 - 0.1 ) ^2 + (0.8-1) ^ 2 + (0 - 0.1) ^ 2 = (0.01 + 0.04+0.01) = 0.06 Not a perfect score but very close to one. XD

Example table..... w/o verifying...

post-102-0-08916400-1306430692.png

There are three categories so you would sum each score in each category... to get the total score for the month.. XD The lower the score the better in this case. :P

^_^;;;

So in summary... I see no issues with decimal forecasts.... ..... :arrowhead:

Link to comment
Share on other sites
Mallow

Mallow

Posted
  • Author
Posted

I've never had a problem with decimal forecast to play the odds for the overall forecast at the end verses trying to score a coop with the exact forecast. XD However this goes back the two semester worth of forecasting contest in my forecasting classes during my time at Penn State. We used a probabilistic based forecast system though in which we put down our confidence in categorical potential values verses laying down a deterministic forecast...

Example in how it would relate to tropical forecast.. (It was graded using a Root Mean squared deviation summation formula... where the lower the the sum of square deviations then the higher the score was ) Threfore a perfect forecast (but risky one.. ) would yield a deviation of 0. Where you forecast a 100 percent (1 :P ) chance of this number happening and it actually verifies. Conversely if you get it wrong..and you did not forecast any probably for the category that actually verified. (say 2 storms form and june... and you said that there was no chance for 2 storms to form in june. Your forcast of 0 percent chance for 2 storms to form would yield a value of 1 for that particular form.

(1-0)^2 = 1 :P You want this value to be 0 XD

more relastic case... For june say this is your probabilistic forecast

0 0.1 0

1 0.8 1

2 0.1 0

(0 - 0.1 ) ^2 + (0.8-1) ^ 2 + (0 - 0.1) ^ 2 = (0.01 + 0.04+0.01) = 0.06 Not a perfect score but very close to one. XD

Example table..... w/o verifying...

There are three categories so you would sum each score in each category... to get the total score for the month.. XD The lower the score the better in this case. :P

^_^;;;

So in summary... I see no issues with decimal forecasts.... ..... :arrowhead:

Cool! :thumbsup:

The main thing I don't like about it is that it's tough to convert the RMSE value to a meaningful value that has a "max score", and where a higher score equals a better forecast. It's much easier with z-scores, since you can assume a normal distribution and convert to a percentile (which fits both criteria).

Also, I doubt people want to put in sixty forecasts (even if a lot of them are zero :P ) for each month.

Link to comment
Share on other sites
Inudaw

Inudaw

Posted
Posted

Cool! :thumbsup:

The main thing I don't like about it is that it's tough to convert the RMSE value to a meaningful value that has a "max score", and where a higher score equals a better forecast. It's much easier with z-scores, since you can assume a normal distribution and convert to a percentile (which fits both criteria).

Also, I doubt people want to put in sixty forecasts (even if a lot of them are zero :P ) for each month.

Lol :XD 0's ftw!

And I don't see why everything has to be "higher number" = better mentality.... but that's only because I pretty much got hooked on the RMSE method on scoring probability based forecast. ^_^;;;

Link to comment
Share on other sites
Mallow

Mallow

Posted
  • Author
Posted

Lol :XD 0's ftw!

And I don't see why everything has to be "higher number" = better mentality.... but that's only because I pretty much got hooked on the RMSE method on scoring probability based forecast. ^_^;;;

It doesn't really matter, of course...

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...