Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

WIP video about Power Laws, kindly requesting feedback! (Patreon)

Published:
2025-11-07 17:50:00
Imported:
2026-09

Videos

  • 561096963.mp4

Downloads

Content

DEADLINE FOR FEEDBACK 10/11 - 12:00 PM UK TIMEWe've got an exciting video coming out very soon!We’d love to hear your feedback and suggestions as we wrap up this video all about power laws!It's a little rough around the edges as we still need to finish a lot animations (as stated in the video) etc. but as always we appreciate you watching and reviewing this cut!Is there anything wrong? Any other questions? Anything more you'd like to see? Explanations you think don't do a good enough job or lack depth? Let us know!

Thank you so much!-Team Ve

Comments

geebeepee

too much, too fast

Charles Bosse

Okay, there are obviously lots of rough edges on this. The things I would say are 1) make sure all the charts are big and easy to read (there is a point with two unreadable sketches presented vertically together) and spend some more time on things that may be mathematically straight forward but hard to parse when you just hear them. "When you take the log of a product, that's the sum of the log" is something I recognize when I see log(A x B) = log(A) + log(B) but struggle with when just "said".

TTST

Delightful! At 21:13, you might reconsider your choice of words to avoid implying that the formation of a large domain would move the system above the Curie temperature. This has many connections to other topics. You mention a few. You could add a connection to prisoners dilemma, in particular spatial PD and its wealth distributions. A widely cited paper about this is https://arxiv.org/pdf/0708.1746

Gregor Shapiro

At 17:30 , attracted to th magnet is not a bolt, it is a (nylock) nut. A steel (iron) bolt (technically, like a rivet does, not have to have threads like a machine screw) would also be attracted to the magnet.

Gregor Shapiro

At 33:00 the animation of the motion of the Earth around the Sun should be more realistic, not today kicked to the Sun (yet), rotation to give all the world day an night, the Moon s rotation and orbit should also be improved. Unfortunately the scale should not be true in this animation the solar system is just too huge, but the barycenter of the three body system could be exaggerated.

Zeus

I don't think it's "too much, too fast". I'd say it's alright, save from some already mentioned technicalities like calling a nut a bolt (grrrr!) or the clunky (but not very important) Sun-Earth animation.

Mirko Serino

1. Ole Peters has proposed a nice time-based resolution of the Saint Petersburg paradox. It does not have directly to do with power laws, but he and his colleagues are building a new brach of economics, which they call (non)-ergodicity economics, which you might want to look at. Not necessarily for this video, but also as an inspiration for future content. https://royalsocietypublishing.org/doi/10.1098/rsta.2011.0065#d1e354 2. When you state that, in games where the average performance is the important parameter, "consistency" is the key...I really would rephrase a bit. If you want to get successful in fat-tailed (power-law-dominated) domains, you need to keep trying even harder then people working with Gaussian systems. Perhaps a nice formulation could be "Consistency is the key for Gaussian/non fat-tailed domains, whereas persistence is the key for domains dominated by power laws." P.S. A NOTE: usually power-law dominated professions are those which have to do with the compounding value of "one-off" information (investing, publishing books of movies...). You can't write the same book twice and nobody can create the same company twice. In food and transportation, what you are offering is more like a commodity. Non across merchants, 'cause everyone provides a different service, but a commodity across time: your food / service should stay consistent. GREAT VIDEO, KEEP UP !

Mike Mitchell

The audio track is way off (lags video).

Mike Mitchell

There is a scene in Black Sheep that goes well with this video. At 35:33, an eagle flies over a boulder and does it's business. That tiny bit of weight hits the boulder, knocking it off it's precarious perch and sends it tumbling down the hill, knocking a cabin off its foundation, where Chris Farley and David Spade are playing checkers. Of course, it's yet another catastrophe in a series of catastrophes set off by unlikely events. That's the stuff of good luck and bad luck.

Rochelle Thompson

4:25 equation should it show “-1.5” currently set as a positive. Really good explanations of the Power laws and how they overflow to other areas. You have been told about the audio but there are also a few spots where there needs to be a cut in the video (these are all in the videos of you not speaking. 18:33, 28:30, 31:55, 41:54, 47:07) Looking forward to seeing the completed vid! Great work!

Stoian Borissov

General feedback: Love the topic and how it connects to your previous videos on network dynamics. I see you covering some ideas from Geoffrey West's book "Scale" as well. I think most of the segments fit well together. I think the self similarity at various scales is very cool and would be nice to see that graphically on the distribution function. Not sure you got a lot of value of the sandpile example. The casino games were nice - though i have a hard time tying it to the following segments on the magnet grains and forest fire simulations. It might be helpful to show a simulation of multiple instances of someone playing the third casino game to demonstrate how it's possible to have very small or giant wins from this highly unstable system. Mistake/Error: 8:23: scale is wrong on y-axis for the binomial log-normal casino game. P(x) should peak at about 0.08, not 0.8 - off by a factor of 10 Comments/Notes: < 6:14: As part of your introduction to different distributions it would help to very briefly tell the audience that there are many different kinds of distributions out there governing different phenomenon and then dive into a few. Before introducing the casino games, it would help to just show a single screen with examples of three different distributions side by side (normal/gaussian, binomial or Poisson, and Pareto/power law), and briefly give examples where each can be found in the world (you already talk about some of this but summarizing it all on a single screen comparing them against each other would help). Then you can turn to the casino games as illustrations of how these distributions work and explain how with each successive example you're going to introduce "fatter" tails. 7:05: "he showed that if you plot the probability of each outcome ..." - which outcome? what was the experiment he was running? need more context. 7:34: I would add that in systems described by normal distributions extreme outliers are highly improbable if not physically impossible - so there are mechanisms which limit the variance of outcomes in a population. For example, as alluded by Dr Strogatz, there are physical, biological, chemical limitations/constraints on how tall a person can be. 12:04: "Distribution of payouts is uncapped" - I get what you're trying to say, and while this is true for this third casino game, the reason as to why it's uncapped compared to the previous games is that it's potentially an infinite game, while the two previous games were defined to be finite (100 trials). I think someone might be mislead to believe that normal distributions are capped due to some mathematical quirk - which is false - as the domain for the gaussian function is +/- inf - the "capping" comes from real world constraints and not constraints on the mathematical function itself. Binomial distributions like that in the second game are only defined for finite number of trials, so they are capped in that sense, however they can be approximated by Poisson distributions as you move towards infinite trials. However, the average (expected value) is undefined for the pareto distribution of the St. Petersburg example, while the average for the normal distribution and binomial distribution is finite and calculable. That fact *IS* due to the inherent equations which govern the distributions. I.e. i could design a theoretical game where the distribution of outcomes is governed by a normal distribution, but the maximum theoretical payout is infinite (while the average is still finite and bounded). 16:20 and 20:06: it would help to demonstrate the self similarity and scale invariance on a graph of the PDF functions itself. Meaning, if you zoom in on just a window of the data (for example on the data from the St. Petersburg paradox) you get the same curve regardless of where you slide your sample window. I did this myself for a sliding window of 9 sequential data points in the St. Petersburg paradox, and it's cool to see how the shape of the curve doesn't change regardless of which 9 points i pick out of list of 35 points i generated. The sliding window in this case is analogous to "zooming in and out" - so the scales of the axes shift, but that's it. This works for both a linear and log-log scale graphs. 33:49: weird clicking sound 42:10: What's also relevant here is that in these examples - the "low performers" are not pruned from the statistics. Because these are real world systems that have scale - they would probably have different looking distributions if you removed poor performers and only looked at the statistics of the top percent of outcomes. In a more resource constrained system, you wouldn't be able to keep lower performers around. 40:20: It's not just different industries - even within the same industries you can get different kinds of distributions depending on the scale of operations. A brick and mortar mom and pop book store will have a different distribution of sales than, let's say Amazon digital book sales. The physical book store faces specific supply chain and physical constraints which limits which books are offered - so it's more likely their sales distributions is more normal. Amazon, on the other hand, has basically zero marginal cost associated with holding tons and tons of books on its digital shelves which only sell a few copies, so it has a huge tail on both ends (a few books that sell a lot, and tons which barely sell at all). Another example would be restaurants. I used to work at a bar with a huge drinks menu - and so when i plotted the inventory of ingredients used in our drinks, the drinks per ingredient graph followed a power law - a few ingredients were used in most of the drinks - while there was a long tail of single use ingredients that were very rarely used. This distribution would look very different if we had a more limited menu selling only a few different drinks. 43:05: Hard to understand what I'm looking at with all the nodes crawling around the edges of the screen/box - only makes sense to me because i saw the previous video on 6 deg of separation. 44:55: Some more comment on how investing and trading is done in this simulation might help contextualize the results. Is it just that winners have more money to invest? Are the bets random? Are some agents better decision makers than others? You say there is no rich getting richer mechanism, so is it just a random walk? Near end/conclusion: To recap, I think it would be helpful to show a single screen comparison showing different kinds of distributions, different example scenarios which are dominated by that behavior, and associated strategy for succeeding in that scenario. Final thoughts: It seems like you are on a streak covering some of the ideas which Geoffrey West talks about in his book "Scale" including concepts regarding power law distributions, fractals, networks, emergent behaviors in complex systems, and scale invariance. If you haven't looked at it already, i think you might get some cool ideas for other ideas you can tie into your current video series.

Beth

Dereks voice is not the voice over voice . This makes it questionable.

Beth

considering his mouth movements are almost in sync........

Beth

Given I trust Derek as an educator.

Timothy-Douglas Alvey

Please introduce people; especially the less “regulars”, like the man in the brown shirt (and if he is from the 1950’s).