Einstein said that "God does not play dice". But perhaps, Fate does?
I have been trying to understand the concept and the philosophy of probability - and whilst I am getting glimmers and glances into the it (and sometimes, I am getting some probable - hmmm - nuggets of wisdom from my readings), I still don't think I get it fully.
Some would probably say, isn't that a bit too odd to study? Or perhaps, argue that "it's too academic - of what use will it be in your work?"
The truth is, I happen to believe that probability - and its related themes: randomization, the search for truth, measurements, errors, risks - are all interesting - and important - stuff that have wide applications.
When one goes to a doctor and presents his symptoms, the doctor actually makes educated guesses based on her experience - which we hope would be deep - and her knowledge - which we hope would be sufficient and inclusive of the latest technologies and advancements. Her prescriptions are based on educated guesses - and that's why there's almost always that advice to "see me 7 days from now if it doesn't improve".
It gives the doctor - and the patient - the chance to "course-correct".
I think at any given time, we are presented with information. Whether those information are perfect or imperfect, complete or incomplete, is immaterial: what we do with those data points and how we scrutinize them in search for nugget of deeper information, insight, knowledge, wisdom - or mere alternate views is what's more critical.
In due time, more information will come in. And that is the time we take corrective action.

This is from _mpd_, from this site. _mpd_ called this "Einstein was wrong".
** And now here comes my diatribe or my discourse on media planning...
One of the missing tools that we have on media planning is risk management - and the basic understanding of measurements. We take Nielsen's or TNS' ratings figures at face value. We assume that the last 13 weeks - and the average of those last 13 weeks - would be a good indicator of the future performance of any certain program. Why 13? Well, because it's always been the case.
We do not take into consideration the variation of the numbers. And the probability of these averages and estimates happening again in the next week, in the week after next, 3 weeks from now, 4 weeks from now. We don't create enough scenarios. We don't apply strict- and rigorous-enough portfolio management principles to media plans.
And let's not get into the "validity" issues: are we measuring program ratings or title ratings, or specific ad/timing- and pagination-rating?
All these need to be taken into account in media planning - and in the calculations of other metrics.
We also do not take enough corrective action. In Singapore (as far as I know and I could be wrong), overnight ratings are not available - data are delivered 48hours/2days (I think - I have been too removed from the buying field). I think overnights would have been ideal - but 2days are still enough to take corrective action against the benchmarks and goals that we have.
Once the plan has been signed and booked and logged into the system, it seems, there is little that one does on the media plan - to take corrective action.
(OK. It just doesn't happen in media planning. It also happens in promotions planning. I have encountered clients who wanted to increase their media investments - "go for more GRPs!" - if the promo doesn't take off. Sure - there is corrective action, but rather, it is a questionable corrective action.)
There just isn't enough thought put into media plans - and promo plans - and marketing plans. There just isn't enough risk management, scenario planning, decision-making disciplines that get into marketing planning.
As the market for consumer goods and services soften in the US - and potentially in other regions - we need to be more aware of the investments - and the associated risks (and probabilities). We owe it to those paying the bills - nope, not the brand managers - the shareholders.