Brings us back to "will not mention name" at the other "will not mention place". Like yourself, I appreciate hard data, i.e., numbers and controlled conditions when testing something and evaluating the results. Science calls for that when it is time to put your cards on the table and making a case using logic - the art and science of sound reasoning. No waffling, little wiggle room, no nuance crap...etc...

Things that need to be avoided are:

Large windows with short intervals in between that almost guarantee success...smaller is better...you made the same point many times at __________....

Large windows scored at 100% followed by a first ring at 90%...should be lot less or nothing at all...else shorten the window...that's a big beef with me...

Explicit rules such as disallowing after a prediction is being argued a miss, "but there was a full moon!!!"...no bringing in a criteria/conditon/observation AFTER ...

I've been thinking more over the last couple hours and find myself requiring two prediction categories...cyclical-based and random/unexpected-based...

Cyclical

24 predictions per year allowed, 2 per month, 3-5 day window with at least 10 days between windows...

To be used by those using tides and syzygy as the driving force behind a prediction attempt. (Anything that comes and goes on a predictable regular/periodic basis.)

Random

50 predictions per year, anytime, any frequency, 3 day windows, but no overlaps. Once they have used up 48 predictions, the prediction form is shut down. The days they are scored against is all the days up the the last day of their last prediction window. I.e, if their 48th prediction expired on Oct 31, then they are scored against 10 months worth of days...

That works out roughly to 150 window days / 300 total days. (Reading my mind Roger?...)

BUT HERE the caveat...they'd be scored against a minimum of 300 days no matter what!!! Those who like to challenge themselves will spread out...and risk a lower score...those who don't and are trying to fish in a barrel will cram their windows together...

A person having a window at the beginning of a year and another that ends on Dec 31. is going to be scored against 365 days...it'll be a toughy for them...it'll require serious contemplation on their part before making a prediction...assuming they're not doing that already...

People using this category are those who driving forces are

Air pressure, water table changes, floods from runoffs, volcanic eruptions, earth generated tones/frequencies, ear tones, headaches, mining depletion operations, 1 million New Zealand sheep going on strike...etc...

Please do suggest alternatives or mods to the random category...

Here's what I'm really thinking...WAY WAY down the road...

At some point Roger, doing the math, you'll discover a cyclical predictor who scores > 60 percent when chance says 50...

At some point Roger, doing the math, you'll discover a random predictor who
scores > 60 percent when chance says 50...

Common sense says, (after verification of course) that you get those two folks together to answer the following:

1. What's the driving force(s) that correlate conversely or inversely between the two successful human from each category...
2. Can those correlating forces be brought together in the form of prediction method that pushes the score toward or past 70 percent success where chance calls for 50 percent?

I'll say it again as I did at _________ many times over...I think it'll require a parallel system twice the size used in weather modelling...then again...there's always lady luck...

=8-)