# Bad predictions – a thing of the past?

Punctuality to time is like precision with distance.

Ages ago I…no, scrub that. I’ll be more precise. On 31 December 2014 I wrote a post about a department mug that each of our team members were given.

Time travel mug shot

“Understanding the past, predicting the future, but it’s coffee now.”

### Weather predictions

At the time I was carrying out research on sea level rise – although it was at a knowledge centre which specialised in weather and climate.

Predicting the weather is difficult! I saw my colleagues using sophisticated numerical models which performed multiple calculations and simulations under a range of scenarios and start conditions, initialised by observations, assumptions and years and years of experience and knowledge.

Were the models successful in making predictions?

Or to rephrase: How confident are we when we hear the weatherman telling us that next Saturday it will be dry at 2 pm when we’re outside cheering our kids on at football?

The reason why weather forecasts appear to be so bad is because we’re very critical on both the spatial and temporal dimensions. An error in either of one them, even to a small degree, means that the forecast (in our opinion) is wrong.

For example, if the rain stops at 14:05 we’ll get wet when the forecast predicted that we’d be dry. Similarly, the rain may come at the predicted time, but at a location 100 meters away – a difference that can be the difference between getting soaked or staying dry.

Making predictions according to our expectations is difficult!

I’m purposefully going to ignore the prediction of the future from tea leaves crap. But one might wonder why we bother with predictive models when the bubbles in our coffee mugs could give an indication of the weather – high atmospheric pressure is associated with fair weather, and in such a case, exerts more pressure on the meniscus the coffee makes in our mug. The visible result is that the bubbles move further to the edges.

Conversely, more bubbles towards the centre of the mug may indicate an increased chance of rain.

### Predictions: The human factor

Whether my ex-colleagues used models, bubbles or pure common sense, some of them seemed to have problems with even the simplest of predictions, such as a moving object in a straight line at constant speed. As an example, a daily occurrence was the blocking of the narrow corridor by people standing in them and chatting.

I’d walk from the coffee machine at one end of the corridor, back to my office at the other end. I needed to squeeze past my gassing colleagues. They’d see me approaching with my mug of tea in my hand, walking at a constant speed in a straight line (recall, this is tea, not beer!) and to a known destination – my office. It couldn’t get any more Newtonian!

And yet it wasn’t until I was standing right next to them they’d figure out that they’d need to move.

So yes. Predicting the future is difficult, and I suppose this is why this aim was emblazoned on our department mugs.

### In sickness and in health

Look what else I found on the (in)side of our department mugs (for clarity, mine is the one of the right!)

I don’t think it needs a genius to predict which one of us is most likely to suffer from sickness. And like the rainfall, the question is where and when?!

### A simultaneous step into the past and future with my new-old mug

You might ask why I’ve dug all of this up from more than 5 years ago.

I’m 2 jobs later. My last job (great company, dreadful job, another mug) is over and I’m now in another great company in a great job. I’m still waiting for my mug though. In the meantime I’ve been using my trusty mug from 2014.

Alas, not so trusty anymore; last week the handle fell off. Somehow that event brought my mind back to these predictive ramblings. Or hindcasts, perhaps.

Now it’s time to move on to a new mug – from my last job. I’m wondering when I’ll experience the present…