Quote from B on June 30, 2022, 1:41 pm

I'd say we are probably nerds so willing to go up in the self-development scale that we have learned (well, I'm still learning it, but I hope to be there soon!) how to be more "normal" than both "text-definition nerds" and "non-nerds" :)

Because let's face it: if Sheldon really applied himself to learn to behave like a normal guy, he would master it so fast!

Quote from Lucio Buffalmano on June 30, 2022, 2:01 pm

Quote from Bel on June 30, 2022, 1:41 pm

Because let's face it: if Sheldon really applied himself to learn to behave like a normal guy, he would master it so fast!

You think?

I personally wouldn't be sure about it.

That's the true "nerd" in the original sense of meaning: with poor social skills, poor social intelligence (potentially or even likely autistic/asperger).
Also not interested in people, while many of us here are interested in people and/or succeeding with people/through people.

I think the Sheldon type of nerd would make huge steps forward, also because he starts from low down, but he'd also be fighting against the odds (and nature).

Quote from B on June 30, 2022, 2:21 pm

Everyone starts where he's at, and "fast" is certainly relative to where one starts from. But as you say, with the exception of really exceptional circumstances, I think one can get better at anything no matter the starting point. And the interest in understanding people, at least to me but I suspect for everybody at some point, inevitably came from wanting to have a better life.

But maybe you're right my joke was just me generalizing my idea of "making it against all odds"... In any case, you already opened my eyes so many times so far, so please don't spoil my dream of Sheldon mastering "normality" Lucio! :)

Quote from leaderoffun on June 30, 2022, 3:44 pm

Quote from Lucio Buffalmano on June 30, 2022, 11:08 am

 

• Naive empiricism (meaningless measure & data points): Ferris likes to measure things, which is a great approach -when possible, producing meaningful measures, and when it's worth it-.
  But he over-measures, including things that's not wort wasting time on, that can't be easily quantified, or where the confounding factors make those measures meaningless.

This is really curious. I get the feeling that we don't measure nearly enough.

Mandatory book reference:

https://www.goodreads.com/book/show/20933591-how-to-measure-anything

Some great quotes:

Anything can be measured. If a thing can be observed in any way at all, it lends itself to some type of measurement method. No matter how “fuzzy” the measurement is, it’s still a measurement if it tells you more than you knew before. And those very things most likely to be seen as immeasurable are, virtually always, solved by relatively simple measurement methods.

Measurement: a quantitatively expressed reduction of uncertainty based on one or more observations.

So a measurement doesn’t have to eliminate uncertainty after all. A mere *reduction* in uncertainty counts as a measurement and possibly can be worth much more than the cost of the measurement.

A problem well stated is a problem half solved.

—Charles Kettering (1876–1958)

## Rule of five

There is a 93.75% chance that the median of a population is between the smallest and largest values in any random sample of five from that population.

Four useful measurement assumptions:

1. Your problem is not as unique as you think.

2. You have more data than you think.

3. You need less data that you think.

4. And adequate amount of new data is more accessible than you think.

I have been a member of two other communities that cared about social skills, and where people posted 'rules' to follow because 'they work'. When you ask them 'how do you know they work?' very rarely the rule creator answered with 'I've tried X times and it worked Y times' or similar.

Once I tried to get everyone to measure the effectiveness of an opening text in Tinder. The belief on that community was that that line was highly effective. When I proposed 'let's just count how often it gets a reply', nobody wanted to do that.

I created a tool to compute stats on Tinder. When I asked people if they wanted to share their data so we could aggregate and have better measurements... nobody did. This was for free. These same people were paying >500 bucks/mo for a community where they receive advice... that they didn't want to test.

So are we self development nerds? Hardly.

Granted, the kind of interactions we learn at PU are hard to replicate IRL (way harder than using an opening text in Tinder or a pickup line IRL). But if you read that book, you DON'T need much for a measure to be effetive. You can measure a lot more than you think.

 

Quote from Lucio Buffalmano on June 30, 2022, 4:26 pm

Great and thought-inspiring post, LOF.

(where's Matthew when you need one, he'd have feasted on this topic :D).

On the "nerds" thing, I switched to thinking we're not (or at least, not because we seek more influence, power, or better relationships).

Now on measuring:

What you say on 93% and median makes sense... When you can properly isolate and measure single variables of interest.

But that isn't always the case:

Quote from leaderoffun on June 30, 2022, 3:44 pm

Anything can be measured. If a thing can be observed in any way at all, it lends itself to some type of measurement method. No matter how “fuzzy” the measurement is, it’s still a measurement if it tells you more than you knew before. And those very things most likely to be seen as immeasurable are, virtually always, solved by relatively simple measurement methods.

That's the crux of the issue.

And that's how I'd quote it instead:

Fuzzy measurements provide a false sense of security and are misleading.

The issue is not so much in measuring and in finding the data -I agree, that's the easy part-.
The issue is what to measure, how to replicate it, how many times... And how to isolate the confounding factors.

When you "measure" single variables in any complex environment you're almost surely setting yourself up for failure.

You :

1. "measure" one thing, you think you're being logical and scienticic
2. maximize around that measure and think you're being effective... And instead, unless you managed to luck it out on measuring the single most significant variable -and to measure it well and in a statistically significant fashion-...
3. but in truth took a random approach instead, wasting time at best, and optimizing around a mistake at worst

 

Example 1: Kolenda and Performance Marketing

I love Kolenda.

But I think I may have lent him too much credibility when it comes to choosing persuasive strategies (something I wanted to tell Ali who quoted him a few times but haven't done it yet).

In theory he's being "scientific".

In practice, his approach is to take single studies, sometimes with a small number of participants, and extrapolates general rules -for example: this color sells more than that, based on this study with 26 people that measured this color against this other 3 colors-.

Then you're supposed to use that "rule" in a bigger page, without any idea of how it interacts with all the rest.
You may be better off than pure chance, but you should still be aware that it's far more of a crapshoot than you think (and you should still measure it again within the bigger page, and with thousands of iterations, if you want to be accurate).

And we're still talking here about "simpler" problems that can be more easily measured.

It gets a lot more complex with more high-level approaches, and with more fluid scenarios:

Example 2: Mehrabian Study

You know the famous quote "communication is only 7% words"?

That's a good example of extrapolating from a study that "measured" things.

So with that measurement you could go your employee and tell him "you're fired".

But if you do so with a smile and positive body language, he should "93% ignore your words".

You see why that doesn't make sense.

Example 3: Small-powered studies are true crapshoots (and they almost all are)

Many times I ran comparison tests for pop-ups.

In theory, you could stop at any number and think that you got a "rough idea".

So if you couldn't go any higher than 100, you'd think "it was a small number, but it's a lot better than no numbers".
And you'd feel confident that the winner is "better".

Yet, when you get the chance of counting in the thousands (rarely the final "winner" started and ended first in my tests), you realize how truly meaningless are the low-powered studies that realy on tends or hundreds or repetition (and almost every single study I've seen in the social sciences stops short of 100).

Quote from leaderoffun on June 30, 2022, 3:44 pm

Measurement: a quantitatively expressed reduction of uncertainty based on one or more observations.

So a measurement doesn’t have to eliminate uncertainty after all. A mere *reduction* in uncertainty counts as a measurement and possibly can be worth much more than the cost of the measurement.

Now, this is a great point.

And I totally agree with you you there.

However, in complex systems that reduction can be truly minimal.

And still, in theory, something is always better than nothing.
Absolutely.

Problem is, the human brain can hardly account for very small variations in probability.
And the "naive empiricism" (an expression borrowed by probability expert Nassim Taleb, BTW) I mentioned earlier does the opposite: it over-extrapolate and jumps to conclusions.

Quote from John Freeman on June 30, 2022, 5:45 pm

When I wrote that I had something else in mind.

I was not thinking about Tim Ferris and the likes.

To use a more positive framing I could have said: “people passionate about self-development” (among those some are nerds as described by you, Lucio).

I mean: not everyone is going on a forum to reflect on their behaviors and interactions. It’s a small minority of the population that is as focused on self-improvement.

In my mind “nerd” was not derogatory but rather affectionate self-deprecating.

Quote from leaderoffun on July 1, 2022, 11:51 am

Oh, I took it as the 'positive' side of nerd! It's just that that involves measuring things, trying to not fool yourself.

And this is... incredibly hard. For every belief we have about social interactions, we should have some amount of evidence that corroborates it. That's the 'nerd' way. Say that PU is a set of say 1000 beliefs about how to interpret social situations and act.

Example (not from PU but from Kavalier's recent streak of blueprints :)

"When someone asks you 'how are you', direct the conversation to something you are currently working on"

The 'method' we have so far is that the community goes 'yep, that nails it'. The more people weighting in on the thread saying 'yep, killed it', the stronger our belief that this is the best way to act becomes.

ll we have a counter to represent how strong this belief is, the counter goes up with opinions of others. Bayesians calls that a prior probability.

Let's say @kavalier post something insightful. I get this 'yep' feeling. My prior goes to .4. Then @amerok posts another 'yep'. Let's say my prior goes to .5 Then Lucio. Because I assign more 'wisdom points' to Lucio, I let my belief go up further when he posts. Now it's .7

Then there are 3 'likes'. my belief goes to.8.

Somehow that belief gets inserted in your identity and you tell people you believe this. You start doing it on social situations.

To be perfectly clear: whether that works or doesn't in social situations has not entered the picture yet; my belief went up to .8 without ANY testing whatsoever. I have zero empirical evidence that this blueprint works (Not picking on you Kavalier, everyone does this all the time, we humans are like this).

Let's say I get a 'how are you', and I respond accordingly with something I'm working on. The conversation flows, and I gain social credit. Boom. My belief goes up to .9. this is the first time empirical evidence enters the picture.

ALL other sources, including your initial 'yep, that should work' feeling (.4) are opinions. Opinions are NOT data, they are opinions. We should not let opinions move our beliefs, or not too much (to see what happens when you don't care about the truth value of what you say or think, look no further: Trump). Twitter is a giant cloud of opinionated people (incentivized to show outrage and extreme opinions) swapping mostly untested opinions and influencing each other.

Back to PU. I put a high prior on each belief I extract from PU because I trust Lucio's criterion. But strictly I don't have evidence that the beliefs are correct till they 'hit reality' and I use them in social situations.

For each of the 1000 beliefs, we should have a 'counter' of how often they work. Of course social situations are anything but stable, and counting  is not easy (lots of ambiguity) but any counting is better than no counting.

I think this is a tremendous effort (clear thinking doesn't come cheap) and most people will not do it. But I do believe it would take the art we are refining here to a whole different level.

Quote from Lucio Buffalmano on July 1, 2022, 12:19 pm

You may call them opinion, but the way many people use it, it's reductive.

And it doesn't give its proper credit.

It almost never is just "opinions".

You may as well call them something like:

Experience-fed and logic-tested results of brain processing power

Sounds much different, no?

That's the power of persuasion and framing (something that came from people's brains BTW more than data).

And it isn't even a manipulative spin, it's what opinions actually are.

Or, at least, that's how good opinions are, the opinions that come with those from experience -ie.: real-world data-, knowledge from study, observation (more data points) and reflection, and a sound brain's logical processing power (ideally supported by solid logic).

Ray Dalio made billions and billions predicting the future with those "opinions".
All fed into a system he called something "believability-weighed" opinions and feeding into a measuring system.

In that system, say, Kavalier's opinion now is not worth 0.4 but 0.6 because he's proven over X number of messages that he delivers great insights.
Lucio's opinion might also be worth something more than 0.4 (because he manages to sound credible :) ).
And a new guy's opinion, or maybe someone who doesn't have the "empowered member" badge -ie.: hasn't gone through PU-, maybe worth 0.1. Maybe even a negative number when he proposes something completely new and he's not supported by anyone else.

In many complex social situations, I trust this system more than I'd trust a single average study.

And it's funny that it seems like I'm attacking science here, while I'm actually a huge supporter and defender of it.

It's not one VS the other, ideally, the three things go together (it's part of TPM's own foundations).

I'm just pointing out the limitations here.

A good approach to learning what works embraces science, but also must be fully aware of its limitations.
And it's important limitations -in the social sciences, especially-.
Overtrusting data without being fully aware of the limitations of any data you're using is what leads to naive empiricism -which is often worse than not relying on any data at all-.

Quote from leaderoffun on July 1, 2022, 1:21 pm

It's not one VS the other, ideally, the three things go together (it's part of TPM's own foundations).

 

Yep I read that post (love it) yesterday as I'm thinking about this problem.

I may be overreacting to my own experience: I noticed I was being influenced (as in adopting a belief) by random opinion from people on twitter/books that don't really deserve much trust. Not that many people are 'clear thinkers' and form their opinion after careful consideration (Lucio does; many others here do; the average twitter/reddit user... not so much).

The really interesting thing is that when you read a book you allow someone to influence you. You have to make a call on whether you want that person to influence you beforehand (hard!). For example, there are books that are clearly a medium for an upsell (a course, or consultancy). There, the author has a clear incentive to make you think in a certain way;  that may or may not have your best interest in mind.

Some books have a clear agenda, but still have a clear, solid helpful bit of wisdom that is useful.

It's damn hard to make the call without reading the book. We have talked about how to identify useful books and why books don't work before. There are some nice tricks there. Including reading the reviews before reading the book (goodreads is great for that).

The question here is: 'how do you value opinion?' Not all opinions are worth the same.

Option 1: ignore opinion by mostly anyone unless they have a verifiable track record OR data (citing sources)

One radical way is to value all opinion at near zero. Read no opinion posts. Don't believe anyone making predictions without checking their track record. Exercise 'writing fact posts' to see the difference between facts and opinions. Example:

https://www.lesswrong.com/posts/Sdx6A6yLByRRs8iLY/fact-posts-how-and-why

The most useful thinking skill I've taught myself, which I think should be more widely practiced, is writing what I call "fact posts."  I write a bunch of these on my blog. (I write fact posts about pregnancy and childbirth here.)

To write a fact post, you start with an empirical question, or a general topic.  Something like "How common are hate crimes?" or "Are epidurals really dangerous?" or "What causes manufacturing job loss?"  

It's okay if this is a topic you know very little about. This is an exercise in original seeing and showing your reasoning, not finding the official last word on a topic or doing the best analysis in the world.

Then you open up a Google doc and start taking notes.

You look for quantitative data from conventionally reliable sources.  CDC data for incidences of diseases and other health risks in the US; WHO data for global health issues; Bureau of Labor Statistics data for US employment; and so on. Published scientific journal articles, especially from reputable journals and large randomized studies.

You explicitly do not look for opinion, even expert opinion. You avoid news, and you're wary of think-tank white papers. You're looking for raw information. You are taking a sola scriptura approach, for better and for worse.

And then you start letting the data show you things. 

You see things that are surprising or odd, and you note that. [continues]

The big downside of this is that it takes a lot of effort to... :

1. Find people who write based on facts (or have some form of wisdom that is impossible to argue with; example personal experience surviving on a raft boat with no drinking water for 49 days)
2. in the topic you are interested in (damn hard, as many topics like social skills don't work that way)
3. find sources yourself (or check the sources provided) and reach your own conclussions)

Option 2: Use prediction markets to understand what the average prediction accuracy is of even accredited experts; follow people with high accuracy only; avoid random tweets or opinion books/posts

Fun fact: even people who are 'famous' and sell their expertise as 'future tellers' DON'T check their predictions. They don't have any estimation of their own accuracy. Neither do their clients. Enter prediction markets. (example: metaculus.com)

Prediction markets enable people to bet on a commodity. In this case, the commodity is a future event, ant participants in the market can bet that the event will or will not occur. They calculate a prediction accuracy per member; some people are better than others.

One downside is that the most popular topics are non-actionable (geopolitics stuff, predictions that are too long term etc)

You can put your predictions there privately, and get an estimate of your own accuracy.

It doesn't work very well for social skills. And it's very far from mainstream. But I'm loving the idea of putting my own businesses assumptions there and calculate my own accuracy on my predictions. I update the predictions as I get more evidence.

Option 3: Read widely, engage with every opinion and evaluate it BEFORE it affects your beliefs. Then it doesn't matter if most of it is crap

This would be ideal,because you don't have to exercise any restraint in what you read.

but in my limited experience, it doesn't work.

You cannot 'unsee' what you have seen.

The mere exposure effect is a thing.

Marketers use this all the time: they hit you with the same idea from multiple angles, and they influence you by sheer brute force: they overwhelm your willpower for critical thinking. They use social cues (well, A has 20k twitter followers and believes this. So do B, C and D, all 'famous'; I'll believe it too)

I often feel that my judgment is clouded after trying to think clearly in front of my computer, mostly reading opinions and having to compute how much I need to update my beliefs. People shouting from their megaphones have perfected the art of influencing you; you the deck is stacked against you.

What other options do you see? I'm sure there are more :)