FAQ: Linear Search: Conceptual - Average Case Performance

Without satellite data, no. Modeling is probabilistic. Even with all the advances we still cannot give more than a one or possibly two day forecast. When it comes to hurricanes the weather forecasters are only able to warn against possible landfall. Even with the wolves at the door, they cannot tell the one by which they gain egress, if at all.

Models give us a way to study a weather event, not predict it. There will be hurricanes, and thunder storms, and wind storms, and ice storms and blizzards. We know that and need only look to the sky to predict weather for our own locale. When a jetliner’s trail stays frozen in the atmosphere for much longer than normal, we know to brace for colder weather. Cold air is aloft which could incubate a snow storm, sleet or freezing rain if there is a warm, moisture laden air mass underneath.

That is exactly what is happening today in our climes. I was walking at 8:30 in snowfall that followed an early morning of rain. It was trying to snow, but the supposed warmth in the air was enough to keep it from happening. None of us can say when or if the temperature will drop by a degree or more that will trigger the transition from water to snow. Bottom line, you don’t want to be on the highway in the dark if that happens while you’re driving.

Computers that can collect data where we stand, and couple in all the other science and empirical data at its disposal is still no better than me taking a walk and looking at the sky or across our valley. If I can see the other side, it’s rain, if not, fog or snow.

2 Likes

Great points, spot on. Can’t argue with evidence. Especially that my standpoint relies pretty heavily on speculation of future computing power and our accuracy.
But is it really so bad? Our weather prediction ability even in the short term is not perfect - but it works extremely well.

Weather forecasting goes a long way back, and thanks to this history we’ve got a great track of accuracy improvement over the years, especially in the last two decades. It can be attributed mainly to the development of new models, pattern study and taking a wider variety of measurements to use in the calculations. The only thing that really changed in 20 years is technology and with it our understanding of the world.

I would argue that this pattern of advancement is universal for all of our simulations and with time, understanding and computing power we will gain higher accuracy of predictions. We need more sensors and more data to generate more accurate thermodynamic models, but science tackles those issues on a daily basis. Mathematicians seem to believe it is a worthwhile task, so it must appear realistically possible. Predictive algorithms such as famous ‘random forest’ and others represent an important branch of data science, with real-life applications. These algorithms are being very intensively developed for purposes of targeted advertisement and marketing, but a collective understanding attained by using these techniques and methods can be applied to other areas of life.

And I didn’t even touch on AI and machine learning algorithms, whose performance relies heavily on the hardware capabilities and whose impact could be tremendous in case of predictive algorithms and our ability to accurately predict seemingly random future events, with use only of real-time data.

That’s great to know. I was worried it’s the sun-dimming particles.

I wish we started this conversation in a different forum… I feel kind of guilty for straying from the topic here, but it’s just too interesting to drop. :frowning:

EDIT: I forgot to include the conclusion - my main point - anything that can be predicted accurately is not random. Considering the trend in our advancements and ability to predict more and more complex model behavior, I am doubtful that anything is truly random and I believe that in time, we will be able to predict with great accuracy any semi-isolated event.

1 Like

i certainly won’t argue that we’re getting closer to making accurate, or reasonably accurate weather predictions. It still depends on visual clues obtained from satellites, ground radar, aircraft and ground observations. Still not completely there, though.

Take for instance tornadoes. In the Prairies we get a few dozen a year, mostly small ones, but some very damaging ones, too. The best the weather forecasters can do is look at the radar and look for intense cells. The where is relatively simple to follow in that respect, but not the when, or the extent of expected damage. We do not know where or if a funnel cloud will form and where or if it will touch down. We can judge speed, and expected intensity, but that’s about all.

Gradually, as models improve weather forecasters might have more to go by.

We are up against an evil giant in terms of Facebook and Google and Apple and Amazon, et al using the technology to target users, especially during elections. The politicians will stop at nothing, and the big guys are more than too eager to help out. The storms that brew from that are the ones that really scare me. Our democracy is crumbling under the weight of misinformation and dubious practices.

1 Like

I understand, these changes are very violent and unstable and occur at a rapid rate, so they must be exponentially more difficult to predict. And yet, at least theoretically, if we know how they form (tornados) and what forces are at play, we should be able to build predictive mathematical models. Perhaps we just don’t understand tornados well enough or we still manage to omit some other important factors involved, so we can’t calculate with practical accuracy.
Similarly to the early weather forecasts, which used mainly only barometer readings as measurements, they weren’t accurate and couldn’t predict much, rather just observe. In time with the inclusion of a wide spectrum of electromagnetic radiation imagery, wind and pressure and other measurments we were able to drastically improve our predictions to the point they can be called fairly reliable.

Unfortunately, I can’t agree more. Let’s hope we’ll survive whatever bubble is going to pop after all this pump of lies has been boiling for long enough.
As bad as it may sound… - Advancement in medical knowledge and human anatomy achieved by Nazis, through conducting horrible and immoral experiments, are the foundation of modern medicine.
I hope at some point we will be able to similarily reclaim all technology developed by the capitalistic machine at our expense and use it for something good and make it accessible to all.

2 Likes