While I can forecast reasonably well I think, my big question I've been thinking about is beyond increasing my score on this website, what tips do people here have on using forecasting to add value to ones life?
I mostly use forecasting to impress my friends and occasionally put a bet on betfair.com. Currently, I have a smallish amount of money on Biden winning. Can't think of anything else.
I really do think there should be a great deal of value in seeing the future before most people do and any tips on how to obtain this value would be really appreciated.
Systematically thinking in probabilities in everyday life helps you make better choices in life, avoid disappointment and see opportunities before others do. Examples from the lifes of people I know: if you know real world statistics about life risks (and they are available about almost everything), you will
- not be completely shocked, if you´re among those 30% whose marriage ends with a divorce (well, if you have serious doubt on the day of your wedding, you might have better cancelled it anyway, because then chances might be much higher)
- have adequate insurance for the really relevant things (what if I can´t work anymore for health reasons (average 20%), die (depends on age and health but is never zero) etc.?) - and no unnecessary ones (does it really make sense to pay for an expensive smartphone insurance, if I can just save the same amount for the next one? etc.))
- can talk your friends out of the newest MLM-scheme, in which only 1 out of 1000 statistically might make real money, and only if the company exists long enough (are they really going to be the one genius with thousands of people under them who sell for an average of 5 bucks a month or might they stay one of the latter forever?)...
- avoid developing a gambling habit (the bank always wins), a drug habit (the harder the stuff, the more probable it is to get hooked) or any other really destructive things
- avoid certain areas, if you like to keep your money, watch, health and life
- avoid certain sports, if you prefer your bones to stay unbroken
etc. etc..
Most successful people are good at unconsciously (very rare) or consciously (can be trained, and very well so on GJO) calculating chance-/risk ratios.
I´m new to this kind of forecasting but have been attaching a probabilty to almost everything for decades already. Really helps to navigate through life in a relaxed way. As said: taking an outside view first for all relevant decisions is a very good habit.
It's hard to point to something specific because the improvements are qualitative. Probabilistic thinking leads to better thinking of risks and benefits, and even being confident in choices where the results turn out poorly because it is a good bet. Most people don't understand that idea, even on this site.
@pstamp Thanks very much for your ideas and thoughts. It's made me (and others) think and is a good start to this discussion.
It's difficult to fully know what life was like before I did forecasting because I think I became a forecaster without having read the book, which means I can't separate easily what I did "before" everything fitted together and what I did after.
But my understanding of most of what you say is essentially probability and not being conned. I remember years ago when all I had was maths skills when I realised gambling was a mugs game. I've never fallen for get rich quick schemes and no doubt am an influence to my friends in that direction. I have prevented a very dangerous scam on a vulnerable person now a friend (though admittedly that was after my law fused with my mathematical skills; was my detection of risk factors combined with my way of extracting the truth of the situation really "forcasting"? Don't know). And I have met other mathematical people equally knowledgable about how silly this gambling and quasi gambling mentality is.
Not blaming myself when a good decision gives an undesired outcome is also something I've done from a very young age.
And I've heard from multiple sources that common advice for insurance is to only insure yourself for what you really need and if you can afford the loss it's worth self-insuring. But again this is probability and / or common sensible risk management that can often be found on the internet quite easily by typing "should I get x insurance" (e.g. https://www.moneyadviceservice.org.uk/en/articles/do-you-need-mobile-phone-insurance).
So it seems that all the benefits you list arise out of understanding probability well and risk management.
If forecasting is just another way of saying "think about probabilities and manage your risks" then I agree, these are the benefits. But are these benefits of some of the skills necessary for forecasting or are they benefits of forecasting itself?
So I agree the route to forecasting leads to acquiring some skills that can have as a bi-product a positive impact on your life. But lets say I have a forecast on something specific and it's good and I'm confident it's correct and most people haven't seen it. I personally have absolutely no idea what to do with it beyond showing off to my friends.
@sebbysteiny I believe that forecasting is indeed mainly about probabilities. Training that ability alone would be worth it. But doing it systematically here leads to some nice others effects as well:
1) You get feedback, hard feedback. If you were wrong, it´s in your brier. No way to get it back out there, no way to tell yourself, "you new the outcame all the time". If you chose that on the last opportunity after having been wrong before, the brier for that question will be bad. In other words: this is a mirror showing the hard way, what is or what you really can, not what one wants to believe.
2) You get an excellent overview of what´s going on in the world. The questions are about almost all topics imaginable. And to answer them correctly, you need to understand the broader context to each of them. So, if you answer many of them, you´re forced to understand the most important aspects of what´s going on outside.
3) You train your ability to do research. You can find most answers about things to know on Wikipedia. So many people believe, that having access to all kinds of information makes them knowledgable. This is about more: knowing and understanding. To predict, if Netanjahu will stay in office for the rest of the year, you need to know some facts about him and his situation (Laws and politics), understand the history of the country and it´s players to a certain degree, approach the question systematically etc.. With blind guessing or only reading one article you end up, where most people are: they have an opinion about everything without having looked at it from all angles, mostly only from the one they´re used to. Only having an opinion and a few facts will definitely lead to bad briers here. That´s a lesson about real life as well, that´s worth a lot. "Winning an argument" might be okay, but it does not automatically mean, that you were right. Perhaps you just had some piece of wrong data, that convinced the other. This won´t work here at all. You need to find the relevant right data and interpret it the right way. You can learn that by studying science or on places like GJO.
4) Another lessons you can also learn when studying science that you get here for free: thinking in categories. To look at a context and systematically analyse all possible outcomes, then attaching probabilities to each of them is even better than attaching probabilities only, because you might miss some things. I found it frustrating to learn that in educational sciences and psychology, because often enough even knowing all possible outcomes it´s nearly impossible to predict human behaviour. Here you only have questions with an outcome that can be measured. Whatever happens - it will be one of the bins. You can aplly that kind of thinking to almost anything usefully, especially in economics.
However good one might be in those aspects: you can always get better through training. Training with a good plan and feedback. If a bodybuilder did not know the weights, he is working with and wouldn´t count the repititions, he couldn´t make systematic progress. Unfortunately too many people don´t apply that principle to thinking.
You won´t build muscles by training here but "brain muscle" and "character muscle". It´s also called Metaskills and can be very uselful in life. And it´s like all other abilities: you can only really understand it, when you have it. It´s definitely a good idea to invest some time here to find out, what you didn´t know before, that you didn´t know. I think, even Einstein could have learned something here ;-)
Your suggestions here seem similar to @pstamp's with less specific examples, and the reply therefore is similar. But you do emphasise more strongly a generic "self-improvement" angle to the probability skills aquired while learning to forecast.
The problem seems to me to be that once you have had that "ah ha" moment, you don't get anything more by continuing to forecast. Or have you found constantly improving forecast skills results in the self-improvement benefits you describe constantly improving also? If so, then in terms of value-add, forecasting is essentially training wheels for the real value of learning to make better decisions in life and there is little "value" in the forecasts themselves, even if you really do know whether the president of Belarus will survive by the end of this year.
It´s not knowing about the president´s fate, that´s in it for you here. It´s far beyound that.
Most questions are cleverly chosen pieces of the bigger picture, just specific enough to be measurable and predictable. With a good forecast on "how many people with Covid-19 will be hospitalized in Texas?" you also say something of the number of Covid deaths there (as most people are hospitalized first and a predictable fraction of them dies). With that you´re saying something about the economic outlook for Texas (opening up or not is depending on those numbers), the employment situation and many other things that might correlate with that. Predicting the number of cars sold in India in Q II tells something about the trend for the whole economy in India, that region or even the whole world. In the last post I forgot one of the most important anspects: thinking in correlations. What is connected to what, not through causality but that way.
Understanding things like that is very useful. In the past months you could really learn tons of lessons about world politics and economy, because that creepy little Virus showed the weaknesses of our globalised world and had consequences on everything.
To sum it up more philosophically a little quote from Ludwig Wittgenstein :
"My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them—as steps—to climb beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend these propositions, and then he will see the world aright."
Adapting that to GJO might look like this:
"What´s in it for you: anyone who understands how to predict the fate of the president of Belarus on a Meta-level eventually recognizes the single question as nonsensical, when he has mastered forecasting them—as steps—to climb beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend the single questions, and then he will see the world aright."
I started forecasting the last year of the first tournament, so 2015. I've never needed to train my research ability because business research in a professional services firm is what I do for a living.
My first year, I mostly forecasted alone. After, I've had the opportunity to do work with other superforecasters. It's a really smart group of people, and when you see how they make their forecasts, it helps build your toolbox. For example, I learned to use R to write functions and automate my forecasts from a PhD student that uses R in their work. It's also clear R is used in evaluations, which can be helpful for assessing overconfidence, etc.
But, I think @pstamp 's point about feedback is probably the key point. You may think you understand probabilistic thinking, but getting feedback helps you to develop your understanding. Personally, I use this platform to try making more radical predictions. When I'm wearing my "super" hat, I try to be much more careful, though there I am much less conservative than most other supers. Which I suppose ties back into the group part, you may think you are smart, but how smart are you compared to the super group, as a whole? That's a really tough benchmark to beat, because it only takes a few screw-ups and you're out. The group doesn't often make the same kind of mistakes individuals do, although it does make its own kind.
I have a tips request.
While I can forecast reasonably well I think, my big question I've been thinking about is beyond increasing my score on this website, what tips do people here have on using forecasting to add value to ones life?
Happy to go first.
I mostly use forecasting to impress my friends and occasionally put a bet on betfair.com. Currently, I have a smallish amount of money on Biden winning. Can't think of anything else.
I really do think there should be a great deal of value in seeing the future before most people do and any tips on how to obtain this value would be really appreciated.
Comment deleted on Sep 08, 2020 03:32PM UTC
Systematically thinking in probabilities in everyday life helps you make better choices in life, avoid disappointment and see opportunities before others do. Examples from the lifes of people I know: if you know real world statistics about life risks (and they are available about almost everything), you will
- not be completely shocked, if you´re among those 30% whose marriage ends with a divorce (well, if you have serious doubt on the day of your wedding, you might have better cancelled it anyway, because then chances might be much higher)
- have adequate insurance for the really relevant things (what if I can´t work anymore for health reasons (average 20%), die (depends on age and health but is never zero) etc.?) - and no unnecessary ones (does it really make sense to pay for an expensive smartphone insurance, if I can just save the same amount for the next one? etc.))
- can talk your friends out of the newest MLM-scheme, in which only 1 out of 1000 statistically might make real money, and only if the company exists long enough (are they really going to be the one genius with thousands of people under them who sell for an average of 5 bucks a month or might they stay one of the latter forever?)...
- avoid developing a gambling habit (the bank always wins), a drug habit (the harder the stuff, the more probable it is to get hooked) or any other really destructive things
- avoid certain areas, if you like to keep your money, watch, health and life
- avoid certain sports, if you prefer your bones to stay unbroken
etc. etc..
Most successful people are good at unconsciously (very rare) or consciously (can be trained, and very well so on GJO) calculating chance-/risk ratios.
I´m new to this kind of forecasting but have been attaching a probabilty to almost everything for decades already. Really helps to navigate through life in a relaxed way. As said: taking an outside view first for all relevant decisions is a very good habit.
It's hard to point to something specific because the improvements are qualitative. Probabilistic thinking leads to better thinking of risks and benefits, and even being confident in choices where the results turn out poorly because it is a good bet. Most people don't understand that idea, even on this site.
Consider doing some personal forecasts e.g. outlined in this post: https://www.lesswrong.com/posts/x4GmqcwjFTnWeRiud/16-types-of-useful-predictions
@pstamp
Thanks very much for your ideas and thoughts. It's made me (and others) think and is a good start to this discussion.
It's difficult to fully know what life was like before I did forecasting because I think I became a forecaster without having read the book, which means I can't separate easily what I did "before" everything fitted together and what I did after.
But my understanding of most of what you say is essentially probability and not being conned. I remember years ago when all I had was maths skills when I realised gambling was a mugs game. I've never fallen for get rich quick schemes and no doubt am an influence to my friends in that direction. I have prevented a very dangerous scam on a vulnerable person now a friend (though admittedly that was after my law fused with my mathematical skills; was my detection of risk factors combined with my way of extracting the truth of the situation really "forcasting"? Don't know). And I have met other mathematical people equally knowledgable about how silly this gambling and quasi gambling mentality is.
Not blaming myself when a good decision gives an undesired outcome is also something I've done from a very young age.
And I've heard from multiple sources that common advice for insurance is to only insure yourself for what you really need and if you can afford the loss it's worth self-insuring. But again this is probability and / or common sensible risk management that can often be found on the internet quite easily by typing "should I get x insurance" (e.g. https://www.moneyadviceservice.org.uk/en/articles/do-you-need-mobile-phone-insurance).
So it seems that all the benefits you list arise out of understanding probability well and risk management.
If forecasting is just another way of saying "think about probabilities and manage your risks" then I agree, these are the benefits. But are these benefits of some of the skills necessary for forecasting or are they benefits of forecasting itself?
So I agree the route to forecasting leads to acquiring some skills that can have as a bi-product a positive impact on your life. But lets say I have a forecast on something specific and it's good and I'm confident it's correct and most people haven't seen it. I personally have absolutely no idea what to do with it beyond showing off to my friends.
@sebbysteiny
I believe that forecasting is indeed mainly about probabilities. Training that ability alone would be worth it. But doing it systematically here leads to some nice others effects as well:
1) You get feedback, hard feedback. If you were wrong, it´s in your brier. No way to get it back out there, no way to tell yourself, "you new the outcame all the time". If you chose that on the last opportunity after having been wrong before, the brier for that question will be bad. In other words: this is a mirror showing the hard way, what is or what you really can, not what one wants to believe.
2) You get an excellent overview of what´s going on in the world. The questions are about almost all topics imaginable. And to answer them correctly, you need to understand the broader context to each of them. So, if you answer many of them, you´re forced to understand the most important aspects of what´s going on outside.
3) You train your ability to do research. You can find most answers about things to know on Wikipedia. So many people believe, that having access to all kinds of information makes them knowledgable. This is about more: knowing and understanding. To predict, if Netanjahu will stay in office for the rest of the year, you need to know some facts about him and his situation (Laws and politics), understand the history of the country and it´s players to a certain degree, approach the question systematically etc.. With blind guessing or only reading one article you end up, where most people are: they have an opinion about everything without having looked at it from all angles, mostly only from the one they´re used to. Only having an opinion and a few facts will definitely lead to bad briers here. That´s a lesson about real life as well, that´s worth a lot. "Winning an argument" might be okay, but it does not automatically mean, that you were right. Perhaps you just had some piece of wrong data, that convinced the other. This won´t work here at all. You need to find the relevant right data and interpret it the right way. You can learn that by studying science or on places like GJO.
4) Another lessons you can also learn when studying science that you get here for free: thinking in categories. To look at a context and systematically analyse all possible outcomes, then attaching probabilities to each of them is even better than attaching probabilities only, because you might miss some things. I found it frustrating to learn that in educational sciences and psychology, because often enough even knowing all possible outcomes it´s nearly impossible to predict human behaviour. Here you only have questions with an outcome that can be measured. Whatever happens - it will be one of the bins. You can aplly that kind of thinking to almost anything usefully, especially in economics.
However good one might be in those aspects: you can always get better through training. Training with a good plan and feedback. If a bodybuilder did not know the weights, he is working with and wouldn´t count the repititions, he couldn´t make systematic progress. Unfortunately too many people don´t apply that principle to thinking.
You won´t build muscles by training here but "brain muscle" and "character muscle". It´s also called Metaskills and can be very uselful in life. And it´s like all other abilities: you can only really understand it, when you have it. It´s definitely a good idea to invest some time here to find out, what you didn´t know before, that you didn´t know. I think, even Einstein could have learned something here ;-)
@cafebedouin
Thanks cafebedoin again for your thoughts.
Your suggestions here seem similar to @pstamp's with less specific examples, and the reply therefore is similar. But you do emphasise more strongly a generic "self-improvement" angle to the probability skills aquired while learning to forecast.
The problem seems to me to be that once you have had that "ah ha" moment, you don't get anything more by continuing to forecast. Or have you found constantly improving forecast skills results in the self-improvement benefits you describe constantly improving also? If so, then in terms of value-add, forecasting is essentially training wheels for the real value of learning to make better decisions in life and there is little "value" in the forecasts themselves, even if you really do know whether the president of Belarus will survive by the end of this year.
It´s not knowing about the president´s fate, that´s in it for you here. It´s far beyound that.
Most questions are cleverly chosen pieces of the bigger picture, just specific enough to be measurable and predictable. With a good forecast on "how many people with Covid-19 will be hospitalized in Texas?" you also say something of the number of Covid deaths there (as most people are hospitalized first and a predictable fraction of them dies). With that you´re saying something about the economic outlook for Texas (opening up or not is depending on those numbers), the employment situation and many other things that might correlate with that. Predicting the number of cars sold in India in Q II tells something about the trend for the whole economy in India, that region or even the whole world. In the last post I forgot one of the most important anspects: thinking in correlations. What is connected to what, not through causality but that way.
Understanding things like that is very useful. In the past months you could really learn tons of lessons about world politics and economy, because that creepy little Virus showed the weaknesses of our globalised world and had consequences on everything.
To sum it up more philosophically a little quote from Ludwig Wittgenstein :
"My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them—as steps—to climb beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend these propositions, and then he will see the world aright."
Adapting that to GJO might look like this:
"What´s in it for you: anyone who understands how to predict the fate of the president of Belarus on a Meta-level eventually recognizes the single question as nonsensical, when he has mastered forecasting them—as steps—to climb beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend the single questions, and then he will see the world aright."
I´m working hard on that one.
I started forecasting the last year of the first tournament, so 2015. I've never needed to train my research ability because business research in a professional services firm is what I do for a living.
My first year, I mostly forecasted alone. After, I've had the opportunity to do work with other superforecasters. It's a really smart group of people, and when you see how they make their forecasts, it helps build your toolbox. For example, I learned to use R to write functions and automate my forecasts from a PhD student that uses R in their work. It's also clear R is used in evaluations, which can be helpful for assessing overconfidence, etc.
But, I think @pstamp 's point about feedback is probably the key point. You may think you understand probabilistic thinking, but getting feedback helps you to develop your understanding. Personally, I use this platform to try making more radical predictions. When I'm wearing my "super" hat, I try to be much more careful, though there I am much less conservative than most other supers. Which I suppose ties back into the group part, you may think you are smart, but how smart are you compared to the super group, as a whole? That's a really tough benchmark to beat, because it only takes a few screw-ups and you're out. The group doesn't often make the same kind of mistakes individuals do, although it does make its own kind.