Tagged: automated trading

Moving Averages

What if we forecasted the weather on temperature moving-averages?

It would seem silly. After all, ENIAC ran the first mathematical computations for a weather model in 1950. ENIAC is not an Icelandic singer. It’s the first true computer and was built by University of Pennsylvania professors in 1946 with funding from the United States Army.

Now, TV weather departments use models that consume data about jet streams, moisture, temperature fluctuations, topography and other factors to project outcomes. For instance, those models say it’ll snow in Denver tonight after 80-degree temperatures the past three days. Chances are they’re right. It’s a significant predictive advance over the old method, the Native American Rock model, in which a rock was hung on a string outdoors. If the rock was wet, it was raining. And so on.

Humans use mathematical models in many predictive ways today. In a subset of weather-forecasting, models anticipate the development and trajectory of hurricanes. We track seismic activity to forecast earthquakes with some measure of warning.

In one of the most interesting applications of mathematical modeling, scientists searching outer space for planets like ours have now identified at least one in a solar system beyond our own. How? With instruments so precise that they can measure differences in light as fine as turning a flashlight on and off on the moon. Slight dimming in measurable light is evidence of a shadow being cast along a path – proof of planets. (more…)