Archives
Blogroll
- Bankelele
- BBC News
- Bossip
- Boston Divorce Lawyer
- Breaking News Kenya
- Business Daily Africa
- Business of Success Podcasts
- coldtusker
- Cool Business Ideas
- CounterPunch
- East African Standard
- Emergency Money Loans
- Entrepreneur Magazine
- Euro Pacific Capital
- Hoodia Green Tea
- Inc Magazine
- Jim Sinclair's Blog
- Jobs in Kenya
- kenyanpoet
- kenyanpundit
- Mish's Global Economic Trend Analysis
- Natural Herbs
- New York Times
- The Daily Nation
- The Nairobi Chronicles
- The Times of India
- Wall Street Gossip
- Wall Street Journal
11 Responses to “Naseem Taleb: Stop Trying to predict the future”
Leave a Reply
Most Comments
- KenCall is not paying it's employee's?! - 597 comments
- Is William Ruto A Dead Man Walking? - 364 comments
- Raila is a cold, calculating, evil man - 146 comments
- Who Are The Richest Kenyans? (Amended) - 139 comments
- I want to talk about the economy - 117 comments
Wangari Maathai – Part 1
What a brilliant woman. So inspiring, intelligent and honorable.
Wangari Maathai – Part 2
Peter Schiff: Listen and Learn
This guy predicted the financial crisis before anyone knew what was happening.
KE,
This guy is really interesting….. There no way of knowing everything…. that’s obvious… But I guess we need to find the optimal point for every context.. and that’s why we need models….
My question is If this guy thinks he knows better Why doesn’t he come up with a better model…. He’s just bullying people who developed models which worked at the time….. If we need to re look at these models then lets do it and move on… NOT spend all the time bickering over randomness… Everything that is uncertain has some randomness in it..
I know u will like this video. Maybe u can research about it and blog about it.
http://www.youtube.com/watch?v.....re=related
Noni:
Yes, I’ve been reading about that story in Dubai. The problem with that emirate, is it’s laws. It’s a crime not to pay your debts in Dubai and so, if you are a foreigner who has now lost your job and can’t afford to pay all your bills, you could potentially end up in jail and that is why they are fleeing.
Also, foreigners have to find a job within one month or their visas expire. That’s another reason why they are fleeing.
One of the many things Dubai will have to do is change those laws, otherwise, those foreigners are not going to return.
However, the arabs in Dubai still have plenty of money. My sister was there recently and she spent a week with one of her workmates. She told me that the amount of money that these Arabs are rolling in, is simply incomprehensible. After leaving Dubai, she landed in Kenya and said that nothing she saw in Kenya and no one she hung out with in kenya, could even come close to the amount of money she saw her friends family spending in Dubai.
YoYo, financial models are not reliable when predicting the future because they are based on past correlations and doubtful assumptions. Economic models are worse “holding all other factors constant; if x, then y???”.
When dealing with future uncertainty always remember Pascals Wager, focus on the consequences of an event which are easily identifiable (then protect yourself from the negative effects) rather than trying to work out the probability of the event which is unknowable. The financial “elite” ignored this and chose to focus on faulty risk models to predict the future…look where that got us!
NNT is not trying to develop alternative models by criticizing current models and just because sth worked in the past doesnt mean it is right, twas just an illusion of validity. To learn more; http://www.fooledbyrandomness.com
Analysans,
My point is there is NO way of knowing everything about the future…
If we knew everything about the future… we’d not have any crises…
Crises is part of how we live here in this world…. Hellow !!!
I agree with the notion that we can try and prevent what negative things we know could happen…. But we still cannot fully know everything negative that what will happen… And financial models are not meant to predict the future.. They’re meant to build a framework to to work with in the present… with the current circumstances… If the variables change.. Then they need to be factored in the models..
This is my point…. You can’t simply wake up in the morning and say quantitative analysis or mean variance is WRONG…… If you say something like this… Then you better show us what you know is RIGHT… You can’t just wake up in the morning and say we need to get rid of a whole school of thought … Just because randomness exists…
This guy is just trying to get ratings AND airtime….. to sell his book…. Which to me is full of pure mysticism…
Anyway that’s my opinion… Randomness is mostly what the whole school of Statistics is about …. So this dude needs to match his skills to what statisticians know about randomness first…. OK.
YoYo:
I believe that financial models were being used to try and predict the future and that’s why the whole thing is now unraveling.
Remember Long-Term Capital Management? I believe one of the founders of that company even won a nobel prize for his bullshit financial model predictors.
And yet, they never learnt.
Bankers are not as smart as they think they are and neither are Ph.d economists. This will be one of the more important lessons that emerges out of this crisis. i.e. bankers are not rocket scientists and they’re going to have to go back to the basics and re-align their salaries with those basic limitations.
KE, LTCM had two “Nobels” Robert Merton and Myron Scholes whose fund had capital base of $5b and accumulated liabilities in excess of $1t.
Yoyo, I agree we cannot know everything about the future yet we have to make decisions about the future bcoz life is the art of making sufficient conclusions from insufficient premises… Models are useful in decision making when what you don’t know about the future may cause a small variation in anticipated outcome, however they cease to be when dealing with certain aspects where what you don’t know may cause a huge variation in outcome.
Statistical analysis is applicable to those aspects where no single observation can determine the total (height, weight, IQ) E.g. if you take a sample of 1000 people, the tallest would make no impact on the total bcoz the randomness is mild. However according to NNT it is not applicable to social sciences and non-linear variables (wealth, income, financial mkts, book sales/author, inflation rates, height btwn species) where wild randomness exists and the total is determined by a small no. of extreme events E.g. what happens when you take a sample of the personal wealth of 1000 people and one of them happens to be Bill Gates or Philip Moi?
Risk models are based on past observations and correlations. Therefore by using them you are implicitly assuming that the future will resemble the past, that investments that produced superior returns in the past will continue to do so in the future. You may experience some initial gains but if a rare event (sub-prime crisis, 9-11, PEV, pandemics, natural catastrophes, unexpected new regulations, innovations by competitors) occurs, you may experience losses that dwarf all your cumulative, previous gains.
NNT does not prescribe alternative models to improve decision making, he instead encourages the use of the enduring ideas we derive from the study of history and the realisation that when trying to grasp/control the future “missing the train is only painful if you run after it!” Therefore, invest in preparedness and not in prediction bcoz chance favours the prepared.
Analyans:
You’re a bit too smart for me.
Can’t you find a way to explain statistics in more simple terms?
non-linear variables, height between species, wha? what?
KE,
I’m very disappointed with you … The reason is because you just jump in to an argument …
You can’t just say all PHD economist and financial experts are not smart or they think they’re smart and they’re not… That a very stupid statement to make especially from you… In the face of future uncertainty nobody is smart… what we work with is too little to call ourselves smart… and this is true all across the board.. in every career that has to deal with future uncertainty !!
Lets talk about the models and the clip NOT about casting aspersions on who is smart and who is not..
**********************
Analysans,
I see where you are going with your point… You see, the only reason we have this type of economic system in the West is because it works with the politics of the day..
NNT major undoing is because he bases his argument on uncertainty(factors we don’t know)…. Genuine scholars do the opposite.. they work from what is known to what is unknown and come up with alternative models…
Anybody can do what NNT is doing .. Its easy to do what he’s doing…. Punching holes in models and theories … By assuming the event of uncertainty..
I just want you to understand this FIRST….
Since we don’t don’t know all the factors that will affect our model – In the event of a change in context..
1.We can only build a model that has the factors that we know…
2. The factors that we don’t know and are going to affect our model can be identified on the basis of how they affect the model
3. If we know how these unknown factors can affect our model then… we can prepare ourselves for these unknown variables…
***********************************
For NNT to just wake up in the morning and start saying we need to get rid of what models we have… this just shows what type of person & scholar he is..
My thing is … he needs to acknowledge first … that the models we work with mostly factor the known… and the rest can only be factored in by a stretch … of probability…
Instead of NNT casting aspersions on Scholars working with what is known… Let him please tell us what Alternatives he has… If he doesn’t then he should just save his breath.. Someone else will find the alternative… He just doesn’t have the guts to.
*************************
KE, just bcoz I subscribe to a view that is anti-conventional wisdom doesn’t make me smarter…I’m old enough to know when am being buttered up. On 2nd thought since I‘m a student I’ll take it as a compliment from a teacher that am on the right track.
Height btwn species (giraffe v cat) will have big variations while height among species will have mild variation, it was just an analogy. I believe my explanation is simple since I tried my best to avoid use of jargon…
Yoyo, working from what is known to what is unknown is ok in normal everyday encounters; bcoz this is how we are wired, it is how the learning process occurs and it saves time. So it is ok to develop general principles over time as we progress in life and rely on the resulting rules of thumb/heuristics to draw conclusions from the new info we encounter.
However when dealing with uncertainty, working from what is known to what is not known may result in a confirmation bias…knowing “if x, then y” may lead you to search for confirming evidence where x & y hold. Instead of the logically correct search for disconfirming evidence, where x & not y hold.
Most models are susceptible to the confirmation bias/illusion of validity since they are based on past observations. Consequently investors who rely on them become overconfident and end up taking bad bets bcoz they fail to realize that they are at an informational disadvantage. For instance, they become over-optimistic about investments that were past winners and later when a correction of the mispricing occurs the past winners under perform.
The current financial crisis (both globally & locally) is partly attributable to the confirmation bias and the resulting overconfidence which cause mkt players to under react to changes in the prevailing trend bcoz they underweight evidence that disconfirms their prior views and overweight any confirming evidence. Consequently they interpret a permanent change as if it were temporary.
“NNT major undoing is because he bases his argument on uncertainty (factors we don’t know)…”
When dealing with the future what you know may be inconsequential bcoz it is what you don’t know that can hurt you. Before 9-11 the authorities knew terrorists can use guns, explosives, biological weapons…and the terrorists knew that they knew, but they did not know that they can use planes.
Please do your homework again, NNT has not stated that we get rid of models…he advices we continue using them where appropriate (physical sciences) but stop relying on them where they are inappropriate (social sciences).
“He advices we continue using them where appropriate (physical sciences) but stop relying on them where they are inappropriate (social sciences)”
It all boils down to having a theoretical model, sufficient empirical observation and statistical analysis. If any of these are not there or inappropriately used, then the approach is not scientifically correct and there may even be no science.
The story of the turkey is amusing, but not very helpful. If you have a model that is confirmed, every day, by the facts, then your model is very useful for prediction of what is going to happen. The turkey story is only helpful for reminding you that you may not have the full context, which will overrule the normal behaviour at some point (Christmas). If the turkey would keep in mind the possibility his owner will kill him, even if he has fed him every day, and based on that runs away from or bites the owner, he will either starve to death, or encourage the owner to kill him long before Christmas. Behaving nicely and eating the food is the best strategy for the turkey, even if Christmas may come.
If social science cannot be based on models and all events in social life are in extremistan, then there is no science. There is just sitting back and waiting for things to happen.
If the models fail, make smarter models. Modelling future events is a core aspect of intelligence.
Me too I found Taleb entertaining, thought provoking but not very useful.