- unofficial guide to forecasting for beginners (will be updated) -
NOTE: There are many people who contributed to this guide. Notably, it contains Philip Tetlock’s commandments, @Anneinak rules and scattered pieces of advice from many other users of this platform. As far as I’m concerned, you’re free to use it in any way you want.
IF YOU WANT TO ADD ANYTHING OR HAVE SOMETHING CHANGED OR REMOVED PLEASE LET ME KNOW.
...
some reasons for forecasting:
understanding, learning
- learning about forecasting
- learning about (some part) of the world/nature
- meta-learning - learning about psychology, statistics, etc. of learning, decision making, thinking, crowd thinking etc.
- testing hypotheses
competition
- competing with other experienced forecasters
- becoming professional forecaster
- earning money with forecasts
leisure
- having fun, playing
- exercising brain
[...]
--------
forecasting advices:
question selection
- triage - focus on questions in the Goldilocks zone of difficulty, where your hard work is likely to pay off
pre-forecast analysis
- complete forecasting principles checklist https://forecastingprinciples.com/index.php/principles-checklist - read each question at least three times, make sure you understand the conditions for resolution and timeframes, ask for clarification if needed
- exercise fermi-izing - break seemingly intractable problems into tractable sub-problems (knowable and unknowable parts)
- conduct creative searches for comparison classes even for seemingly unique events
- look for the clashing causal forces at work in each problem (know the power players (or forces) and their preferences (nature)
- know the norms and protocols of institutions
- remember about black swans, wild cards, accidents, exponentials, fat tail distributions etc.
- resist the lure of wishful thinking (ask yourself: are you forecasting or making a wish?)
- think like them (decision makers)
- identify possible outcomes
- list, in advance, the signs that would nudge you toward the other argument
- strike the right balance between inside and outside views.
- strive to distinguish as many degrees of doubt as the problem permits but no more (also remember that everything should be made as simple as possible but not simpler)
- strike the right balance between under- and overconfidence, between prudence and decisiveness.
- if possible, use mathematical and statistical models (all models are wrong but some are useful)
- hunt for the right information (google, twitter, newspapers, etc., set notifications for keywords), seek out diversity of sources
- compare your forecast beforehand with that of the experts in the field, other good forecasters, the crowd, prediction markets etc.
- weigh your forecast against your level of knowledge and expertise in the matter - in general the less you know, the less confident and closer to 50% you should be
- assess credibility of sources, recognize the different levels of trustworthiness for evidence
- try to understand the process
- remember that correlation does not imply causation
- know the statistics of the class to which the case belongs
- avoid unnecessary assumptions, logical fallacies and biases - debias your forecast as much as possible (self-reflection and evidence-based reasoning is the cure to bias)
- remember that past performance is no guarantee of future results
- remember about regression toward the mean and the time frame of the question
- look for rules of a thumb in various types of phenomena and processes, e.g. the status quo is the usual outcome
- be careful and conservative with questions involving high degree of randomness and/or fluctuation, e.g. sports, stocks etc.
- look (primarily) for disconfirming evidence
- analyze your attitude and extent of your knowledge
- don’t assume you could make a calculated mathematical expectation for every forecast; sometimes you have to take a risk and move forward with the information at hand, even if it’s incomplete - intuition and judgment are necessary evils in an incalculable world.
- know when to stop (https://en.wikipedia.org/wiki/Optimal_stopping)
- find your own rules of a thumb in searching, analyzing etc.
- avoid perfection, look for optimal solutions instead
writing the rationale, commenting
- treat your rationale as the forecast
- your rationale acts as a memory aid
- your rationale enables backtesting (https://en.wikipedia.org/wiki/Backtesting)
- writing long and honest rationale allows you to see the entire logic from a birds eye view and permits people to poke holes in your reasoning
updating
- master the fine art of bayesian belief updating
- strike the right balance between under- and overreacting to evidence
- tease subtle signals from noisy news flows
- look for harbingers, usually a different person (journalist, forecaster etc.) will catch some piece of news or discover a pattern
- be willing to admit error and change course (when the facts change, change your mind)
post-forecast analysis
- don’t forget to do postmortems
- keep a notebook (or a document on your device) to track and analyze your forecasts
- look for the errors behind your mistakes but beware of rearview-mirror hindsight biases
- look for patterns and recurring themes in your mistakes
- look for strategies that were successful
cooperation
- bring out the best in others and let others bring out the best in you
- master the fine art of perspective taking (understanding the arguments of the other side so well that you can reproduce them to the other’s satisfaction)
- master the fine art of precision questioning (helping others to clarify their arguments so they are not misunderstood)
- master the fine art of constructive confrontation (learning to disagree without being disagreeable)
- orient the conversation toward truth seeking
- seek to enrich others with your knowledge and experiences and avoid harm, insult at all cost
- in discussions, state your assumptions clearly, avoid vague or ambiguous verbiage, reframe if necessary
- let others be wrong if they choose to, end conversations when they become fruitless
- don’t waste the precious time of other forecasters, comment concisely when you have something new and valuable
- don’t bring propaganda and ideology to the platform
- remember that like eating has chewing, speaking (and writing) has thinking
competition
- outwork other forecasters, research more, analyze more, update more
- work smarter
- forecast early for accuracy (or late for brier score), you may want to wait 2-10 days to think and wait for the consensus to settle down
- the crowd is wrong half the time, the trouble is figuring out which half
- don’t flip, but if you really need to, go to 50% first
- when an event happens, think carefully, does the event impact your forecast?
- make relatively many forecasts for accuracy (or relatively few for brier score)
- if you get more advanced and well-calibrated it may be useful to extremize at least some of your forecasts (i.e. go 0% or 100%)
- one unappreciated way you could get extraordinary scores is by looking for a question, where you have knowledge others can’t access. You can propose such questions here https://www.gjopen.com/questions/1687-what-forecasting-questions-should-we-ask-what-questions-would-you-like-to-forecast-on-gjopen .
- be very careful with going polar (0 or 100%) and/or strongly against the crowd consensus if you don’t have complete and unambiguous evidence, especially if your forecast diverges greatly from other good forecasters
- make forecasts similar to experts/good forecasters, for brier and accuracy scores in general it’s better to be often wrong slightly than wrong completely once
forecasting tools
- write down your forecast-making process
- make mind maps, decision trees or anything that could enhance your thinking and decision/forecast-making process
- learn some R or python
news reading
- in questions about events taking place in other countries read local news if possible
- in questions susceptible to bias (e.g. politics, ideology) read news from both sides of the spectrum (don’t forget that most, if not all, newspapers are usually skewed toward one side of the spectrum in their reporting, identify their stance - abridgenews.com)
- set notifications
- learn to recognize fake news and distinguish opinions from facts
progressing
- identify your weaknesses
- stay humble and growth-oriented forever
- believe it’s possible to get better (knowledge and practice are far more important than talent)
- set specific, measurable, attainable, relevant and time-bound goals with your forecasting practice
- progressively overload (don’t take too many or too difficult forecasts, but also beware of always answering easy and similar questions if you want to learn and progress)
- take your time to adapt
- learn from the best
- be determined, consistent and diligent, keep at it as long as it takes to get your desired results
- master the error-balancing bicycle. (practice, practice, practice and keep score of your forecasting progress)
- don't get so caught up in Brier score, in fact forget about your Brier in your learning phase
- all forecasts are wrong, but some are useful
- to become an expert, make all the mistakes which can be made in a very narrow field
- borrow ideas from various other disciplines
- never fear making mistakes, err and err and err again, but less and less and less
- take time to analyze your past forecasts
- check your thinking for biases and recurring mistakes
- value diverse views, always learn from others (even beginners or people who are wrong most of the time)
- search constantly for valuable information - read, listen and think about forecasting
- look for a new law by the following process: first guess it, then compute the consequences of the guess to see, if this law you guess is right, what it would imply and then compare the computation results to observations to see if it works. If it disagrees with experiment, it’s wrong
- strive to be cautious, humble, nondeterministic, actively open-minded, curious, intelligent, reflective, numerate, pragmatic, analytical, dragonfly-eyed, probabilistic
biases in forecasting
- read Thinking, Fast and Slow by D. Kahneman
- if short on time, scan some list of cognitive biases and logical fallacies, e.g. this - http://humanhow.com/list-of-cognitive-biases-with-examples/ and this https://www.logicallyfallacious.com/logicalfallacies/search [...]
--------
some major spectra of forecasting approaches:
- hedgehog-like (closed-minded, having unshakable self-confidence, with one/few big idea(s)) - fox-like (open-minded, treating beliefs as hypotheses, drawing on diverse ideas)
- sharing (bringing out the best in others and letting others bring out the best in you) - quiet (never commenting)
[...]
--------
some major case studies of forecasting approaches:
NOTE: if I misrepresented someone's approach or you want yourself (or someone else) off or on this list or just want to add anything please do comment, I will make sure to add/amend/remove promptly
- textbook superforecasting
example: @Anneinak;
also: @bobrown, @fisherdbus [...]
my comment: If you want to see textbook superforecasting characteristics and principles in practice, I recommend following @Anneinak (or anyone she recommends) and studying her approach in resolved forecasts. It's probably the safest approach for beginners to follow.
- extremizing
example: @Wildwood2000;
also: [...]
my comment: Obviously, @Wildwood2000 has a low Brier score with his rather unique approach, but I would caution beginners not to make the mistake of mimicking him, because most likely you’re not going to get the same results, because a) he’s more experienced than you, in this approach it's crucial to be right most of the time b) as far as I know, he’s doing some extensive research before going polar c) mistakes hurt his Brier much less, if you’re a beginner you’d be blown out
- playing with consensus
example: ;
also: [...]
my comment: I may be wrong, but from what I’ve seen I suspect that many of the most successful forecasters (those at the top of the leaderboards) are playing with consensus, that is they forecast many questions against the crowd as opposed to giving their opinions.
- wild guessing:
example: @slobodan;
also: [...]
my comment: This forecaster's approach is truly remarkable, I recommend studying his forecasts, especially toward the beginning, he was at the top of the leaderboards in quite many questions, often when the crowd was completely wrong. Yet his brier score and challenge rankings ended up worse than most. I don't know anything about his approach, whether I’m correct to assume he was wild guessing and was simply lucky. Or perhaps he knew something others didn’t (maybe he’ll enlighten us). Takeaway here is you may be very right in many questions yet get very poor results in the long run.
[...]
NOTE: The goal in studying various approaches of other forecasters is to learn and enrich your own approach. It's not wise to ignore someone's approach, even if it's wrong most of the time, each approach has it's own advantages and disadvantages. Diversity is strength.
--------
glossary of terms related to forecasting
https://forecastingprinciples.com/dictionary/defined%20terms.html for now refer to wikipedia
if you have any questions or want to add some definitions please comment
--------
some major sources on forecasting:
books
primary
P. Tetlock, D. Gardner, Superforecasting: The Art and Science of Prediction
N. Silver, The Signal and the Noise
R.J. Hyndman, G. Athanasopoulos, Forecasting: Principles and Practice (https://otexts.com/fpp2/)
R. Hogarth, Judgement and Choice
[...]
Dear @Mi,
I must agree with @Jayce, we would learn much more if you shared your approach with us, me guessing would be pretty useless.
I hoped listing some examples of approaches would start a brainstorming discussion, where anyone could add something to this guide.
I encourage all fellow forecasters to share their own views, it can be an approach, advice or anything which could be added to this guide. I want this to be a collective work, the more diverse the better.
About sharing and competing, I’d like to quote @cdob63: “I am okay with telling everybody everything I know, which is usually not much, because I think of this exercise as a construction project--we are all working together to sort out the Big Picture--and not a contest. This is not as oleaginous and sanctimonious as it probably sounds--several decades of experience have suggested to me that I do better in Whatever It Is when everybody does better.”
And also recommend this article: https://www.alfiekohn.org/article/contest-2/
@pstamp Thanks nonetheless, this @K-M-G \ User Name has nothing to do with me, unless he or she just follows me blindly LOL, I have been the only forecaster going against the Crowd in a few,.
*
My system is not that Complicated
*
LOL
*
It’s not just what you forecast it’s how you forecast
Baldwin (2016) January 15, states that:
The Five Traits
1 Superforecasters begin by gathering as much information as possible.
2 Superforecasters nurture and develop the habit of thinking in terms of probability when exploring the likelihood of a specific event
3 Forecasting improves when individuals work in groups
4 Superforecasters ensure that they are regularly keeping score of their projections
5 The most successful forecasters are willing to admit error and quickly change course on their projections.
@sebbysteiny stated that:
\
"I see you are very likely to use 0% or 100% guesses".
\
Your score would improve if you did not guess, you research and analyzes the data and make an informed assessment of probability. Point in case, if you think I guess, flip a coin 400 times and see how many times you call it right.
\
If you can time after time call 340/358 Your the Man.
@K-M-G thank you for sharing this guide. You gathered a lot of wisdom and experience here.
I am happy that you have found my Spotify list of forecasting podcasts episodes worth mentioning in your recommended resource list.
I also have my youtube list of videos related to forecasting: https://www.youtube.com/playlist?list=PLnUdbSv_SNUu--30vjEDQC5d26-12s0DC (it might be a little chaotic at the moment but I add to that list any good videos about forecasting that I see on youtube)
As you know I would like to promote forecasting in Poland. To do that I am working on a free youtube course about the introduction to forecasting in Polish. I have the help of my friend Alex (alexrjl on Metaculus) who is a much more experience forecaster than me. Alex is progressing with his own youtube Introduction to forecasting course: https://twitter.com/lxrjl/status/1323368883407986688 and I try to help him with that however I can. If you have time and interest, I would be very happy to show you my first video when it is finished. With your experience and understanding of Polish, it would be great to know your thoughts and suggestions.
- unofficial guide to forecasting for beginners (will be updated) -
NOTE: There are many people who contributed to this guide. Notably, it contains Philip Tetlock’s commandments, @Anneinak rules and scattered pieces of advice from many other users of this platform. As far as I’m concerned, you’re free to use it in any way you want.
IF YOU WANT TO ADD ANYTHING OR HAVE SOMETHING CHANGED OR REMOVED PLEASE LET ME KNOW.
...
some reasons for forecasting:
understanding, learning
- learning about forecasting
- learning about (some part) of the world/nature
- meta-learning - learning about psychology, statistics, etc. of learning, decision making, thinking, crowd thinking etc.
- testing hypotheses
competition
- competing with other experienced forecasters
- becoming professional forecaster
- earning money with forecasts
leisure
- having fun, playing
- exercising brain
[...]
--------
forecasting advices:
question selection
- triage - focus on questions in the Goldilocks zone of difficulty, where your hard work is likely to pay off
pre-forecast analysis
- complete forecasting principles checklist https://forecastingprinciples.com/index.php/principles-checklist
- read each question at least three times, make sure you understand the conditions for resolution and timeframes, ask for clarification if needed
- exercise fermi-izing - break seemingly intractable problems into tractable sub-problems (knowable and unknowable parts)
- conduct creative searches for comparison classes even for seemingly unique events
- look for the clashing causal forces at work in each problem (know the power players (or forces) and their preferences (nature)
- know the norms and protocols of institutions
- remember about black swans, wild cards, accidents, exponentials, fat tail distributions etc.
- resist the lure of wishful thinking (ask yourself: are you forecasting or making a wish?)
- think like them (decision makers)
- identify possible outcomes
- list, in advance, the signs that would nudge you toward the other argument
- strike the right balance between inside and outside views.
- strive to distinguish as many degrees of doubt as the problem permits but no more (also remember that everything should be made as simple as possible but not simpler)
- strike the right balance between under- and overconfidence, between prudence and decisiveness.
- if possible, use mathematical and statistical models (all models are wrong but some are useful)
- hunt for the right information (google, twitter, newspapers, etc., set notifications for keywords), seek out diversity of sources
- compare your forecast beforehand with that of the experts in the field, other good forecasters, the crowd, prediction markets etc.
- weigh your forecast against your level of knowledge and expertise in the matter - in general the less you know, the less confident and closer to 50% you should be
- assess credibility of sources, recognize the different levels of trustworthiness for evidence
- try to understand the process
- remember that correlation does not imply causation
- know the statistics of the class to which the case belongs
- avoid unnecessary assumptions, logical fallacies and biases - debias your forecast as much as possible (self-reflection and evidence-based reasoning is the cure to bias)
- remember that past performance is no guarantee of future results
- remember about regression toward the mean and the time frame of the question
- look for rules of a thumb in various types of phenomena and processes, e.g. the status quo is the usual outcome
- be careful and conservative with questions involving high degree of randomness and/or fluctuation, e.g. sports, stocks etc.
- look (primarily) for disconfirming evidence
- analyze your attitude and extent of your knowledge
- don’t assume you could make a calculated mathematical expectation for every forecast; sometimes you have to take a risk and move forward with the information at hand, even if it’s incomplete - intuition and judgment are necessary evils in an incalculable world.
- know when to stop (https://en.wikipedia.org/wiki/Optimal_stopping)
- find your own rules of a thumb in searching, analyzing etc.
- avoid perfection, look for optimal solutions instead
writing the rationale, commenting
- treat your rationale as the forecast
- your rationale acts as a memory aid
- your rationale enables backtesting (https://en.wikipedia.org/wiki/Backtesting)
- writing long and honest rationale allows you to see the entire logic from a birds eye view and permits people to poke holes in your reasoning
updating
- master the fine art of bayesian belief updating
- strike the right balance between under- and overreacting to evidence
- tease subtle signals from noisy news flows
- look for harbingers, usually a different person (journalist, forecaster etc.) will catch some piece of news or discover a pattern
- be willing to admit error and change course (when the facts change, change your mind)
post-forecast analysis
- don’t forget to do postmortems
- keep a notebook (or a document on your device) to track and analyze your forecasts
- look for the errors behind your mistakes but beware of rearview-mirror hindsight biases
- look for patterns and recurring themes in your mistakes
- look for strategies that were successful
cooperation
- bring out the best in others and let others bring out the best in you
- master the fine art of perspective taking (understanding the arguments of the other side so well that you can reproduce them to the other’s satisfaction)
- master the fine art of precision questioning (helping others to clarify their arguments so they are not misunderstood)
- master the fine art of constructive confrontation (learning to disagree without being disagreeable)
- orient the conversation toward truth seeking
- seek to enrich others with your knowledge and experiences and avoid harm, insult at all cost
- in discussions, state your assumptions clearly, avoid vague or ambiguous verbiage, reframe if necessary
- let others be wrong if they choose to, end conversations when they become fruitless
- don’t waste the precious time of other forecasters, comment concisely when you have something new and valuable
- don’t bring propaganda and ideology to the platform
- remember that like eating has chewing, speaking (and writing) has thinking
competition
- outwork other forecasters, research more, analyze more, update more
- work smarter
- forecast early for accuracy (or late for brier score), you may want to wait 2-10 days to think and wait for the consensus to settle down
- the crowd is wrong half the time, the trouble is figuring out which half
- don’t flip, but if you really need to, go to 50% first
- when an event happens, think carefully, does the event impact your forecast?
- make relatively many forecasts for accuracy (or relatively few for brier score)
- if you get more advanced and well-calibrated it may be useful to extremize at least some of your forecasts (i.e. go 0% or 100%)
- one unappreciated way you could get extraordinary scores is by looking for a question, where you have knowledge others can’t access. You can propose such questions here https://www.gjopen.com/questions/1687-what-forecasting-questions-should-we-ask-what-questions-would-you-like-to-forecast-on-gjopen .
- be very careful with going polar (0 or 100%) and/or strongly against the crowd consensus if you don’t have complete and unambiguous evidence, especially if your forecast diverges greatly from other good forecasters
- make forecasts similar to experts/good forecasters, for brier and accuracy scores in general it’s better to be often wrong slightly than wrong completely once
forecasting tools
- write down your forecast-making process
- make mind maps, decision trees or anything that could enhance your thinking and decision/forecast-making process
- learn some R or python
news reading
- in questions about events taking place in other countries read local news if possible
- in questions susceptible to bias (e.g. politics, ideology) read news from both sides of the spectrum (don’t forget that most, if not all, newspapers are usually skewed toward one side of the spectrum in their reporting, identify their stance - abridgenews.com)
- set notifications
- learn to recognize fake news and distinguish opinions from facts
progressing
- identify your weaknesses
- stay humble and growth-oriented forever
- believe it’s possible to get better (knowledge and practice are far more important than talent)
- set specific, measurable, attainable, relevant and time-bound goals with your forecasting practice
- progressively overload (don’t take too many or too difficult forecasts, but also beware of always answering easy and similar questions if you want to learn and progress)
- take your time to adapt
- learn from the best
- be determined, consistent and diligent, keep at it as long as it takes to get your desired results
- master the error-balancing bicycle. (practice, practice, practice and keep score of your forecasting progress)
- don't get so caught up in Brier score, in fact forget about your Brier in your learning phase
- all forecasts are wrong, but some are useful
- to become an expert, make all the mistakes which can be made in a very narrow field
- borrow ideas from various other disciplines
- never fear making mistakes, err and err and err again, but less and less and less
- take time to analyze your past forecasts
- check your thinking for biases and recurring mistakes
- value diverse views, always learn from others (even beginners or people who are wrong most of the time)
- search constantly for valuable information - read, listen and think about forecasting
- look for a new law by the following process: first guess it, then compute the consequences of the guess to see, if this law you guess is right, what it would imply and then compare the computation results to observations to see if it works. If it disagrees with experiment, it’s wrong
- strive to be cautious, humble, nondeterministic, actively open-minded, curious, intelligent, reflective, numerate, pragmatic, analytical, dragonfly-eyed, probabilistic
biases in forecasting
- read Thinking, Fast and Slow by D. Kahneman
- if short on time, scan some list of cognitive biases and logical fallacies, e.g. this - http://humanhow.com/list-of-cognitive-biases-with-examples/ and this https://www.logicallyfallacious.com/logicalfallacies/search
[...]
--------
some major spectra of forecasting approaches:
- hedgehog-like (closed-minded, having unshakable self-confidence, with one/few big idea(s)) - fox-like (open-minded, treating beliefs as hypotheses, drawing on diverse ideas)
- conservative (risk-averse, luck-independent) - bold (risk-taking, luck-dependent)
- intuitive (relying on intuition, heuristics) - quantitative (relying on models, firm statistical inference)
- independent (ignoring most or all of the other forecasters) - crowd-tracking (relying on the wisdom of the crowd)
- extensive (breadth-first, jack-of-all trades) - intensive (depth-first, master of one)
- sharing (bringing out the best in others and letting others bring out the best in you) - quiet (never commenting)
[...]
--------
some major case studies of forecasting approaches:
NOTE: if I misrepresented someone's approach or you want yourself (or someone else) off or on this list or just want to add anything please do comment, I will make sure to add/amend/remove promptly
- textbook superforecasting
example: @Anneinak;
also: @bobrown, @fisherdbus [...]
my comment: If you want to see textbook superforecasting characteristics and principles in practice, I recommend following @Anneinak (or anyone she recommends) and studying her approach in resolved forecasts. It's probably the safest approach for beginners to follow.
- extremizing
example: @Wildwood2000;
also: [...]
my comment: Obviously, @Wildwood2000 has a low Brier score with his rather unique approach, but I would caution beginners not to make the mistake of mimicking him, because most likely you’re not going to get the same results, because a) he’s more experienced than you, in this approach it's crucial to be right most of the time b) as far as I know, he’s doing some extensive research before going polar c) mistakes hurt his Brier much less, if you’re a beginner you’d be blown out
- playing with consensus
example: ;
also: [...]
my comment: I may be wrong, but from what I’ve seen I suspect that many of the most successful forecasters (those at the top of the leaderboards) are playing with consensus, that is they forecast many questions against the crowd as opposed to giving their opinions.
- understanding the process
example: @sebbysteiny;
also: @SE_Meyer, @ManuPV [...]
- wild guessing:
example: @slobodan;
also: [...]
my comment: This forecaster's approach is truly remarkable, I recommend studying his forecasts, especially toward the beginning, he was at the top of the leaderboards in quite many questions, often when the crowd was completely wrong. Yet his brier score and challenge rankings ended up worse than most. I don't know anything about his approach, whether I’m correct to assume he was wild guessing and was simply lucky. Or perhaps he knew something others didn’t (maybe he’ll enlighten us). Takeaway here is you may be very right in many questions yet get very poor results in the long run.
[...]
NOTE: The goal in studying various approaches of other forecasters is to learn and enrich your own approach. It's not wise to ignore someone's approach, even if it's wrong most of the time, each approach has it's own advantages and disadvantages. Diversity is strength.
--------
glossary of terms related to forecasting
https://forecastingprinciples.com/dictionary/defined%20terms.html
for now refer to wikipedia
if you have any questions or want to add some definitions please comment
--------
some major sources on forecasting:
books
primary
P. Tetlock, D. Gardner, Superforecasting: The Art and Science of Prediction
N. Silver, The Signal and the Noise
R.J. Hyndman, G. Athanasopoulos, Forecasting: Principles and Practice (https://otexts.com/fpp2/)
R. Hogarth, Judgement and Choice
[...]
secondary
https://research-doc.credit-suisse.com/docView?language=ENG&format=PDF&source_id=csplusresearchcp&document_id=1065113751&serialid=Z1zrAAt3OJhElh4iwIYc9JHmliTCIARGu75f0b5s4bc%3D
D. Kahneman, Thinking, Fast and Slow
M. Pirie, How to Win Every Argument
D. Berry, Statistics, A Bayesian Perspective
J. Baron, Thinking and Deciding
P. Boghossian, J. Lindsay, How to Have Impossible Conversations, A Very Practical Guide
edge.org annual question series
[...]
videos and podcasts
https://www.youtube.com/playlist?list=PLnUdbSv_SNUu--30vjEDQC5d26-12s0DC
https://open.spotify.com/playlist/4LKES4QcFNozmwImjHWrBX
https://open.spotify.com/show/7DXTunzwdz4mFrvajWrhQ4
https://braininspired.co/podcast/
[...]
gjopen, comments:
https://www.gjopen.com/comments/1124259
https://www.gjopen.com/comments/1127067
https://www.gjopen.com/training
articles and websites
https://forecastingprinciples.com
https://philosophy.hku.hk/think/
http://changingminds.org/disciplines/argument/fallacies/fallacies_alpha.htm
https://www.logicallyfallacious.com/logicalfallacies/search
http://humanhow.com/list-of-cognitive-biases-with-examples/
https://www.mpib-berlin.mpg.de/staff/gerd-gigerenzer/publications?page=1&fbclid=IwAR0ZigU_tx4QqgahQDUEJsN8jv2ZhaWBMrNVWC2ZuEB0_4VlkiMFPepB0Wk
https://ocw.mit.edu/courses/find-by-topic/#cat=mathematics&subcat=probabilityandstatistics
https://powerful-problem-solving.com/ideas/ and https://powerful-problem-solving.com/resources/
https://dawlab.princeton.edu/publications/?fbclid=IwAR1GmpU507i2FzXuJ8h0vGQmTYwUI3P8znPXiWmSiEeS9P_bxT2a62z8ZmA
https://fbri.vtc.vt.edu/content/fbri_vtc_vt_edu/en/people-directory/primary-faculty/montague/publications.html
[...]
twitter
https://twitter.com/annieduke/status/1313653673994514432?s=21
Dear @Mi,
I must agree with @Jayce, we would learn much more if you shared your approach with us, me guessing would be pretty useless.
I hoped listing some examples of approaches would start a brainstorming discussion, where anyone could add something to this guide.
I encourage all fellow forecasters to share their own views, it can be an approach, advice or anything which could be added to this guide. I want this to be a collective work, the more diverse the better.
About sharing and competing, I’d like to quote @cdob63: “I am okay with telling everybody everything I know, which is usually not much, because I think of this exercise as a construction project--we are all working together to sort out the Big Picture--and not a contest. This is not as oleaginous and sanctimonious as it probably sounds--several decades of experience have suggested to me that I do better in Whatever It Is when everybody does better.”
And also recommend this article: https://www.alfiekohn.org/article/contest-2/
@pstamp Thanks nonetheless, this @K-M-G \ User Name has nothing to do with me, unless he or she just follows me blindly LOL, I have been the only forecaster going against the Crowd in a few,.
*
My system is not that Complicated
*
LOL
*
It’s not just what you forecast it’s how you forecast
Baldwin (2016) January 15, states that:
The Five Traits
1 Superforecasters begin by gathering as much information as possible.
2 Superforecasters nurture and develop the habit of thinking in terms of probability when exploring the likelihood of a specific event
3 Forecasting improves when individuals work in groups
4 Superforecasters ensure that they are regularly keeping score of their projections
5 The most successful forecasters are willing to admit error and quickly change course on their projections.
Baldwin, G. (2016). It’s not just what you forecast, it’s how you forecast. Futures.
Retrieved from
http://www.futuresmag.com/2016/01/15/it%E2%80%99s-not-just-what-you-forecast-it%E2%80%99s-how-you-forecast
@sebbysteiny stated that:
\
"I see you are very likely to use 0% or 100% guesses".
\
Your score would improve if you did not guess, you research and analyzes the data and make an informed assessment of probability. Point in case, if you think I guess, flip a coin 400 times and see how many times you call it right.
\
If you can time after time call 340/358 Your the Man.
@Wildwood2000
"I see you are very likely to use 0% or 100% guesses" [followed by stuff about guessing]
Thanks for highlighting the typo. By "guess" I meant "forecasts" or "answers", now corrected.
I attached no substantive significance to the word "guess" but all feedback is welcome if it helps make my posts clearer.
I'll try to comment later.
@Chrisp Are they still blocking most of your posted rationales?
@K-M-G thank you for sharing this guide. You gathered a lot of wisdom and experience here.
I am happy that you have found my Spotify list of forecasting podcasts episodes worth mentioning in your recommended resource list.
I also have my youtube list of videos related to forecasting: https://www.youtube.com/playlist?list=PLnUdbSv_SNUu--30vjEDQC5d26-12s0DC (it might be a little chaotic at the moment but I add to that list any good videos about forecasting that I see on youtube)
As you know I would like to promote forecasting in Poland. To do that I am working on a free youtube course about the introduction to forecasting in Polish. I have the help of my friend Alex (alexrjl on Metaculus) who is a much more experience forecaster than me. Alex is progressing with his own youtube Introduction to forecasting course: https://twitter.com/lxrjl/status/1323368883407986688 and I try to help him with that however I can. If you have time and interest, I would be very happy to show you my first video when it is finished. With your experience and understanding of Polish, it would be great to know your thoughts and suggestions.
@michal_dubrawski,
I'm not that experienced and still am experimenting here, but sure I look forward to sharing my views and helping in any way I can.
@K-M-G thank you very much, that is very generous of you - can I contact you directly in any way? My email is in the Facebook section on my profile.