markov_user

• markov_userApr 2, 2021, 12:42 PM

  30 points

  0 ∶ 0

  on: An­nounc­ing “Nam­ing What We Can”!

  I believe there are a few serious flaws in your guesstimate model:

  • a year has 365.2421905 days, not 365.25. That’s not even rounded correctly!

  • Smiles per QALY should multiply days in a year with smiles in a good day, instead they are added. They don’t even have the same unit, how can you add them! Insanity!

  • the post’s karma is far outside of even your 99% interval

  Everything else seems quite correct and I agree with your CIs and conclusions.

  Also, please find a new name for guesstimate.

• markov_userApr 2, 2021, 12:19 PM

  1 point

  0 ∶ 0

  on: Statis­tics for Lazy Peo­ple, Part 1

  Nice post! Found it through the forum digest newsletter. Interestingly I knew Lindy’s Law as the “Copernican principle” from Algorithms to Live By, IIRC. Searching for the term yields quite different results however, so I wonder what the connection is.

  Also, I believe your webcomic example is missing a “1 -”. You seem to have calculcated p(no further webcomic will be released this year) rather than p(there will be another webcomic this year). Increasing the time frame should increase the probability, but given the formula in the example, the probability would in fact decrease over time.

• markov_userJan 30, 2021, 9:52 AM

  2 points

  0 ∶ 0

  on: EA Mün­ster Pre­dic­tions 2021

  “Bei 80% der Treffen der EA Münster Lokalgruppe in 2021 waren mehr als 5 Personen anwesend”—how will cancelled meetups (due to lack of attendees, if that ever happens) count into this? Not at all, or as <=5 attendees? (kind of reminds me of how the Deutsche Bahn decided to not count cancelled trains as delayed)

  Also, coming from EA Bonn where our average attendance is ~4 people, I find the implications of this question impressive. :D

• markov_userJan 2, 2021, 11:39 AM

  2 points

  0 ∶ 0

  in reply to: MaxRa’s comment on: One’s Fu­ture Be­hav­ior as a Do­main of Calibration

  I see, so at the end of the day you’re assigning a number representing how productive the day was, and you consider predicting that number the day before? I guess in case that rating is based on your feeling about the day as opposed to more objectively predefined criteria, the “predictions affect outcomes” issue might indeed be a bit larger here than described in the post, as in this case the prediction would potentially not only affect your behavior, but also the rating itself, so it could have an effect of decoupling the metric from reality to a degree.

  If you end up doing this, I’d be very interested in how things go. May I message you in a month or so?

• markov_userJan 2, 2021, 11:30 AM

  2 points

  0 ∶ 0

  in reply to: MaxRa’s comment on: One’s Fu­ture Be­hav­ior as a Do­main of Calibration

  Good point, I also make predictions about quarterly goals (which I update twice a month) as well as my plans for the year. I find the latter especially difficult, as quite a lot can change within a year including my perspective on and priority of the goals. For short term goals you basically only need to predict to what degree you will act in accordance with your preferences, whereas for longer term goals you also need to take potential changes of your preferences into account.

  It does appear to me that calibration can differ between the different time frames. I seem to be well calibrated regarding weekly plans, decently calibrated on the quarter level, and probably less so on the year level (I don’t yet have any data for the latter). Admittedly that weakens the “calibration can be achieved quickly in this domain” to a degree, as calibrating on “behavior over the next year” might still take a year or two to significantly improve.

• markov_userDec 31, 2020, 3:50 PM

  3 points

  0 ∶ 0

  on: One’s Fu­ture Be­hav­ior as a Do­main of Calibration

  I personally tend to stick to the following system:

  • Every Monday morning I plan my week, usually collecting anything between 20 and 50 tasks I’d like to get done that week (this planning step usually takes me ~20 minutes)

    • Most such tasks are clear enough that I don’t need to specify any further definition of done; examples would be “publish a post in the EA forum”, “work 3 hours on project X”, “water the plants” or “attend my local group’s EA social” – very little “wiggle room” or risk of not knowing whether any of these evaluates to true or false in the end

    • In a few cases, I do need to specify in greater detail what it means for the task to be done; e.g. “tidy up bedroom” isn’t very concrete, and I thus either timebox it or add a less ambiguous evaluation criterion

  • Then I go through my predictions from the week before and evaluate them based on which items are crossed off my weekly to do list (~3 minutes)

    • “Evaluate” at first only means writing a 1 or a 0 in my spreadsheet next to the predicted probability

    • There are rare exceptions where I drop individual predictions entirely due to inability to evaluate them properly, e.g. because the criterion seemed clear during planning, but it later turned out I had failed to take some aspect or event into consideration^[1], or because I deliberately decided to not do the task for unforeseeable reasons^[2]. Of course I could invest more time into bulletproofing my predictions to prevent such cases altogether, but my impression is that it wouldn’t be worth the effort.

  • After that I check my performance of that week as well as of the most recent 250 predictions (~2 minutes)

    • For the week itself, I usually only compare the expected value (sum of probabilities) with actually resolved tasks, to check for general over- or underconfidence, as there aren’t enough predictions to evaluate individual percentage ranges

    • For the most recent 250 predictions I check my calibration by having the predictions sorted into probability ranges of 0..9%, 10..19%, … 90..99%.^[3] and checking how much the average outcome ratio of each category deviates from the average of predictions in that range. This is just a quick visual check, which lets me know in which percentage range I tend to be far off.

    • I try to use both these results in order to adjust my predictions for the upcoming week in the next step

  • Finally I assign probabilities to all the tasks. I keep this list of predictions hidden from myself throughout the following week in order to minimize the undesired effect of my predictions affecting my behavior (~5 minutes)

    • These predictions are very much System 1 based and any single prediction usually takes no more than a few seconds.

    • I can’t remember how difficult this was when I started this system ~1.5 years ago, but by now coming up with probabilities feels highly natural and I differentiate between things being e.g. 81% likely or 83% likely without the distinction feeling arbitrary.

    • Depending on how striking the results from the evaluation steps were, I slightly adjust the intuitively generated numbers. This also happens intuitively as opposed to following some formal mathematical process.

  While this may sound complex when explaining it, I added the time estimates to the list above in order to demonstrate that all of these steps are pretty quick and easy. Spending these 10 minutes^[4] each week seems like a fair price for the benefits it brings.

  ---

  1. ↩︎

     An example would be “make check up appointment with my dentist”, but when calling during the week realizing the dentist is on vacation and no appointment can be made; given there’s no time pressure and I prefer making an appointment there later to calling a different dentist, the task itself was not achieved, yet my behavior was as desired; as there are arguments to be made to evaluate this both as true or false, I often just drop such cases entirely from my evaluation

  2. ↩︎

     I once had the task “sign up for library membership” on my list, but then during the week realized that membership was more expensive than I had thought, and thus decided to drop that goal; here too, you could either argue “the goal is concluded” (no todo remains open at the end of the week) or “I failed the task” (as I didn’t do the formulated action), so I usually ignore those cases instead of evaluating them arbitrarily

  3. ↩︎

     One could argue that a 5% and a 95% prediction should really end up in the same bucket, as they entail the same level of certainty; my experience with this particular forecasting domain however is that the symmetry implied by this argument is not necessarily given here. The category of things you’re very likely to do seems highly different in nature from the category of things you’re very unlikely to do. This lack of symmetry can also be observed in the fact that 90% predictions are ~10x more frequent for me in this domain than 10% predictions.

  4. ↩︎

     It’s 30 minutes total, but the first 20 are just the planning process itself, whereas the 3+2+5 afterwards are the actual forecasting & calibration training.

One’s Fu­ture Be­hav­ior as a Do­main of Calibration

• markov_userDec 30, 2020, 12:19 PM

  1 point

  0 ∶ 0

  on: An­nounc­ing the Fore­cast­ing In­no­va­tion Prize

  “Before January 1st” in any particular time zone? I’ll probably (85%) publish something within the next ~32h at the time of writing this comment. In case you’re based in e.g. Australia or Asia that might then be January 1st already. Hope that still qualifies. :)

• markov_userDec 20, 2020, 4:10 PM

  1 point

  0 ∶ 0

  in reply to: Neel Nanda’s comment on: Make a Public Com­mit­ment to Writ­ing EA Fo­rum Posts

  Indeed, thank you. :) I haven’t started the other, forecasting related one, but intend to spend some time on it next week and hopefully come up with something publishable before the end of the year.

• markov_userDec 14, 2020, 8:48 AM

  3 points

  0 ∶ 0

  on: CFAR Work­shop in Hindsight

  My thoughts on how to best prepare for the workshop (as mentioned in the post):

  • Write down your expectations, i.e. what you personally hope to take away from the workshop (and if you’re fancy, maybe even add quantifications/​probability estimates to each point)

  • Make sure you can go into the workshop with a clear head and without any distractions

  • Don’t make the same mistake I made, which was booking a flight home way too early on the day after the end of the workshop. I didn’t realize beforehand how difficult it was to get from the workshop venue to the airport, and figuring out a solution stressed me quite a bit during the week (but was in the end solved for me by the super kind ops people)

  • Do your best in the week(s) before to stay healthy

  • Sleep enough the nights before

  • Maybe prepare a bug list and take it with you; this will also be one of the first sessions, but the more the better

  • Don’t panic; if you don’t manage to prepare in any significant way, the workshop is still extremely well designed and you’ll do just fine.

CFAR Work­shop in Hindsight

• markov_userNov 22, 2020, 10:14 AM

  7 points

  0 ∶ 0

  in reply to: NunoSempere’s comment on: An­nounc­ing the Fore­cast­ing In­no­va­tion Prize

  Sure. Those I can mention without providing too much context:

  • calibrating on one’s future behavior by making a large amount of systematic predictions on a weekly basis

  • utilizing quantitative predictions in the process of setting goals and making plans

  • not prediction-related, but another thing your post triggered: applying the “game jam principle” (developing a complete video game in a very short amount of time, such as 48 hours) to EA forum posts and thus trying to get from idea to published post within a single day; because I realized writing a forum post is (for me, and a few others I’ve spoken to) often a multi-week-to-month endeavour, and it doesn’t have to be that way, plus there are surely diminishing returns to the amount of polishing you put into it

  If anybody actually ends up planning to write a post on any of these, feel free to let me know so I’ll make sure focus on something else.

• markov_userNov 20, 2020, 4:30 PM

  15 points

  0 ∶ 0

  on: Make a Public Com­mit­ment to Writ­ing EA Fo­rum Posts

  Good timing and great idea. Considering I’ve just read this: https://​​forum.effectivealtruism.org/​​posts/​​8Nwy3tX2WnDDSTRoi/​​announcing-the-forecasting-innovation-prize I’ll gladly commit to submitting at least one forum post to the forecasting innovation prize (precise topic remains to be determined). Which entails writing and publishing a post here or on lesswrong before the end of the year.

  I further commit to publishing a second post (which I’d already been writing on for a while) before the end of the year.

  If anybody would like to hold me accountable, feel free to contact me around December 20th and be very disappointed if I haven’t published a single post by then.

  Thanks for the prompt Neel!

• markov_userNov 20, 2020, 4:24 PM

  3 points

  0 ∶ 0

  on: An­nounc­ing the Fore­cast­ing In­no­va­tion Prize

  Nice! In the few minutes of reading this post I came up with five ideas for related things I could (and maybe should) write a post on. My only issue is that there’s only 6 weeks of time for this, and I’m not sure if that’ll be enough for me to finish even one given my current schedule. But I’ll see what I can do. May even be the right kind of pressure, as otherwise I’d surely be following Parkinson’s law and work on a post for way too long.

  (The many examples you posted were very helpful by the way, as without them I would have assumed I don’t have much to contribute here)

• markov_userNov 13, 2020, 4:37 PM

  6 points

  0 ∶ 0

  on: Have you ever used a Fermi calcu­la­tion to make a per­sonal ca­reer de­ci­sion?

  Firstly, Monte Carlo simulations (such as on https://​​www.getguesstimate.com/​​ ) is likely more precise/​​useful than pure Fermi estimates, as they make uncertainty, including your consideration of the value possibly being off by factor 10, explicit and thus you can have greater confidence in the results. One advantage Fermi estimates definitely have is that they force you to think about the different components of the problem, or in this case how different careers contribute to your impact. But they are generally speaking primarily helpful to estimate orders of magnitude, and are thus not all that useful in comparing different options unless they lie very far apart.

  Secondly, I agree with shawzach that it makes a lot of sense to talk your career considerations through with other people. I used EAGx Virtual for that purpose and looked for all the people who might have something to contribute to my career considerations. But then again, it differed from your situation in that I didn’t have several alternatives but rather one preexisting plan and wanted to figure out whether it was “good” or I should look further for alternatives. Still, people might be able to add a few crucial considerations or just arguments you hadn’t thought of before that affect your estimates.

• markov_userOct 27, 2020, 11:47 AM

  1 point

  0 ∶ 0

  on: The Ve­gan Value Asym­me­try and its Consequences

  But in order to have net positive lives, we need to do something more than follow consumer-choice based principles.

  I agree. Veganism is (for most vegans, I believe) mostly about reducing the harm you inflict on the world. It’s clear you can’t ever get to 0. Even if your life is net positive, somewhere along the way you always harm somebody or some being. And while veganism itself certainly has this asymmetry you refer to, it seems a lot of vegans take steps beyond that in the more positive direction, such as

  • being effective altruists and in that way trying to do more good in the world

  • go into activism or animal rights advocacy

  • work at animal shelters or even just taking care of a stray animal

  So I don’t think the risk of neglecting the positive side of things is all that high. Certainly makes sense to take it into consideration though, and I appreciate your post!

• markov_userSep 26, 2020, 7:11 AM

  7 points

  0 ∶ 0

  in reply to: Prabhat Soni’s comment on: Prab­hat Soni’s Shortform

  I don’t have or know of any data (which doesn’t mean much, to be fair), but my hunch would be that rationalist people who haven’t heard of EA are, on average, probably more open to EA ideas than the average altruistic person who hasn’t heard of EA. While altruistic people might generally agree with the core ideas, they may be less likely to actually apply them to their actions.

  It’s a vague claim though, and I make these assumption because of the few dozens of EAs I know personally, I’d very roughly assume ²⁄₃ of them to come across as more rationalist than altruistic (if you had to choose which of the two they are), plus I’d further assume that from the general population more people will appear to be altruistic, than rationalist. If rationalists are more rare in the general population, yet more common among EAs, that would seem like evidence for them being a better match so to speak. These are all just guesses however without much to back them up, so I too would be interest in what other people think (or know).

• markov_userSep 14, 2020, 10:54 AM

  1 point

  0 ∶ 0

  on: A few re­ally quick ideas about per­sonal finance

  But use with caution, as I think the time/​$ trade-off might imply you should maximise your working hours, which I think isn’t a good assumption. See point earlier about knowledge workers.

  I definitely agree with “use with caution”. The value of my time question has been bothering me for years now, and I never really found a satisfying heuristic for it. There are just a lot of complications and there imho is no way this can reasonably be considered to be a constant value:

  • it depends not only on the amount of time you’re thinking about spending in a certain way (or gaining), but on the difference of value of the thing you’re considering you do as compared to the thing you would be doing otherwise

  • even if you manage to pin it down to some value in a situation, that’s only the marginal value of time given your circumstances, and it would differ if you had more/​less time or money on your hands; thus, if you derive some great heuristic from that (e.g. “take a cab after parties”), depending on how much that impacts your behavior, this changes the value of your time, so the heuristic is sort of self-defeating to a degree

  So I basically came up with a range of the value certain things I tend to do are worth per hour, and in that way have a way to compare things. It does feel like I’m overthinking and under-utilizing the principle though… one conclusion it led me to however was that I should work less, since my time was worth more to me than what I was paid per hour (after taxes). This is definitely a case where point 2 from above is highly relevant though. Reducing my time at the job rather drastically changes the value of my time, so I need to take that into account and try to find an equilibrium in working just as much (or little) such that the estimated value of my time (maybe averaged over my typical daily habits) reduces to roughly what I’m being paid to work. Which is tough!

  But yeah, I’m getting a bit rambly. Just one more thing: Consumption smoothing is an interesting concept which I have to admit never occurred to me before. Thanks for the post! It also nudged me to once again look into investing which I’ve been procrastinating for years.

• markov_user 5 Jul 2020 19:40 UTC

  4 points

  0 ∶ 0

  on: Con­sider a wider range of jobs, paths and prob­lems if you want to im­prove the long-term future

  Designing recommender systems at top tech firms

  Semi-related and somewhat off-topic, so forgive me for following that different track – but I recently thought about how one of the major benefits of EAGx Virtual for me was that it worked as a recommender system of sorts, in the form of “people reading my Grip profile (or public messages in Slack) and letting me know of other people and projects that I might be interested in”. A lot of “oh you’re interested in X? Have you heard of Y and Z?” which often enough led me to new interesting discoveries.

  I’m curious if there may be a better approach to this, rather than “have a bunch of people get together and spontaneously connect each other with facts/​ideas/​people/​projects based on mostly random interactions”. This current way seems to work quite well, but it is also pretty non-systematic, luck-based, and doesn’t scalethat well (it kind of does, but only in the “more people participate and invest time → more people benefit” kind of way).

  (That all being said, conferences obviously have a lot of other benefits than this recommender system aspect; so I’m not really asking whether there are ways to improve conferences, but rather whether there are different/​separate approaches to connecting people with the information most relevant to them)

• markov_user 19 Jun 2020 17:14 UTC

  11 points

  0 ∶ 0

  on: effek­tiv-spenden.org: 2019 - Year in Review

  Just a random experience report—I’ve been using the website for my monthly donations for about half a year now, and think it’s great. The process is so easy and friction-less that donating is something I’m always looking forward to as opposed to it feeling like an obligation I need to get through with, which basically had been my feeling about it beforehand. The website’s great UX really makes a huge difference.

  Also it’s a great and easy page to point people to that sympathize with the idea of effective giving but don’t really know what the next steps would be. Which isn’t surprising as I guess that’s partly the reason the project exists in the first place.

  Random thought—I’ve seen many EAs apply stickers (e.g. GiveWell) to their phones or laptops, which I guess is a very cheap tool that may lead to a couple of conversations (and an expected >0 “conversions”) about EA or the respective org in particular. I’d assume that at least marginally such stickers are pretty effective little things, but it may depend on how difficult it is to set up and distribute such a product. So it may certainly not be on top of your backlog if anything, but have you put any thought into such ideas? Maybe something like this would be a good opportunity for a volunteer, as I imagine it would require very little coordination initially.