Anyone else feel like most of the playerbase has quit at this point?
2 Likes
I donât feel like theyâve quit the game. I do think theyâve quit visiting these forums and honestly that might be for the best. As far as the game goes, weâll also see if having it on GamePass gives it a jump for a little while.
5 Likes
Well anytime I log into the game I play the same 5 people, last month it was the same 10. Just seems like the game is basically dead at this point.
If you play at the same time of day every day, especially when itâs night or early morning for most people on the server, unsurprising
1 Like
Load your replays in scelight and multi-analyze them. This opens a complete list of all unique players (excluding smurfs). You ctrl a ctrl c ctrl v into google sheets and that tells you the number of players for a given time period if you sort the sheet by date of last play. In the past 11 months Iâve had 894 unique opponents. My most common opponent Iâve only played 30 times which is 1 game every 10 days. Average games vs all opponents is 4. So in a yearâs time youâll play the same person maybe 4 times in a row. And thatâs assuming you play as much as I do, which is very unlikely â Iâve averaged 8 games a day.
Your account has averaged 6.8 games/day which is 85% of mine. Extrapolating, that means you run into your average opponent ~4 games/year.
This is a manifestation of the BaaderâMeinhof phenomenon which is also known as frequency illusion. Your brain automatically tracks things it cares about & ignores things it doesnât care about, which makes you over-estimate how frequently things happen. Thatâs what happens w/ sc2 opponents: you think you run into the same person a bunch, but you really donât. Your brain is rigged in a thousand ways to prevent a proper statistical inference from occuring, which is why researchers go to such extreme lengths as randomizing and double-blinding studies.
Statistics and probability theory are so counter intuitive because the human brain is rigged to not be able to understand them. This is so severe that even I, a math savant from the age of 10, only scored an 88 (of 99) on probability meanwhile I scored 99 in every other category at that age. Itâs really hard to do properly and this is so extreme that entire fields of PhD graduates still mess it up. So donât beat yourself up over it. If you want an example of this happening in action, look no farther than the computer models used for climate science. They are horrible models that are the incarnation of bias, but they are touted as absolute proof that Earth is doomed unless we adopt communism. The researchers simply donât understand the statistical foundation of their own models and their heads explode if you explain it to them.
Here is a doctorate physicist explaining some of the obvious problems:
So she is basically saying the data shows the models simply arenât accurate and that this inaccuracy is underestimated because itâs the average of many models without incorporating each modelâs own individual inaccuracy (they assume the individual model inaccuracy is 0, when we know thatâs false because the models donât agree with one another so the inaccuracy canât be 0). So if you incorporate individual model inaccuracies, the accuracy goes down even more. But the issue is thatâs itâs impossible to know the accuracy of each individual model. Translation, they under-estimate the inaccuracy and the accuracy is still really bad even so.
She doesnât explain why this is the case, but itâs pretty obvious. What they are simulating isnât actual water. Water has parameters that make it impossible to model, itâs an unsolved problem in physics, so they simulate a type of fluid that isnât water. They make it weakly compressible for example. On a baseline, the simulations arenât doing the same thing as reality. So of course they arenât going to be good at making predictions. Imagine modeling what a baseball will do but you use the parameters of a bowling ball and then say âOMG why was my prediction so far off?â and thatâs basically what climate scientists do.
Frankly, this is the biggest issue that mankind faces right now because it underpins every other issue. The inability to analyze data without bias is an absolutely monstrous problem that affects every other problem on Earth. Everything from Ubisoft going bust to balance in SC2 being rigged for protoss to climate science â itâs all downstream from the issue of being unable to apply statistics to interpret data.
Why do you bother with the AI posts? Not intelligent enough to come up with your own reply?
2 Likes
Chat GPT outscores 3rd-year medical students on clinical reasoning exams. Itâs smart, but nowhere near smart enough to criticize an entire industry filled with doctoral physicists without being laughed out of the room. Climate models are an intersection of applied physics and computer science, and the physicists donât recognize the limitations of their models due to that intersection. That happens because they donât understand computer science. Ramsey theory dictates that any sufficiently complex process can be tuned to output any pattern. Furthermore, you can find an infinite number of processes capable of outputting a given pattern. When you have a complex model outputting a pattern that visually resembles some data, all youâve done is create a glorified regression analysis program. You could get a better fit using more basic regression analysis techniques. They are adding more assumptions and getting worse results. Outputting a pattern means your algorithm is robust, not that it is physically accurate.
For an algorithm to be physically accurate, it must be derived from 1st order principles. You must predict how the climate behaves based on how water molecules behave on an individual level. This is fundamentally impossible due to their sheer quantity of the particles and you canât treat the particles as a group because fluids are chaotic. Itâs literally the definition of an unsolvable problem as per Godelâs incompleteness theorems. Even if you could, in theory, build a super computer capable of modelling every molecule of water, you couldnât measure the initial state of the waterâs configuration without disturbing its state via the measurement. So you canât derive the answer from 1st order principles and that means your algorithm will always be subject to ramsey theory.
It gets worse. The measurement problem means that models similar to quantum uncertainty apply as well. Now you gotta incorporate the Schrodinger equation. Lmao. Good luck.
haha, your reply makes zero sense in relation to what you actually replied to. Donât you read the Chat GPT garbage you post before you post it?
I played the same person 5 times in a row in 1 session. Your math isnât mathing.
He isnât smart enough to realize the AI garbage he posts makes zero sense.
Average. For every 1 person you play 30 times/year, youâre running into 8-9 players a singular time.
Can solve with linear equation.
(30 + N) / (N + 1) = 4
N=8.7
Top term is total games. Bottom term is number of players. That calculates average games per player, which must equal 4. Granted this is a linear approximation assuming singular game opponents, but thereâs a spectrum of players and some you will face more and some you will face less. That means you need statistics and calculus to solve this issue because you need the ratio of the integrals for the probability density function for the probability distribution that models the frequency of games played. Then you could say you play â10 people 15 times for every one person you play 7 timesâ or whatever. But the average is all that matters because it tells you the vast majority of your opponents you will face 1 time and a small minority you will face 50 times.
In fact, itâs a pareto distribution:
https://i.imgur.com/dybDDs2.png
Here is a much better chart:
https://i.imgur.com/4Vyhhg9.png
460 single-time opponents. 190 two-time opponents. Etc.
1 Like
Can I get a recipe for a good ratatouille?
I personally never had Ratatouille, but it is one of the listed meals in Babishâs channel:
https://youtu.be/s_ymnJmtvMs?si=UcaO0UpGo6QGq5DT&t=20
Looks quite nice too for final result.
My apologies. That was soup.
This is ratatouille or something of it.
https://www.youtube.com/watch?v=roCX0AfBseQ
The game is almost 15 years old, of course most of the player base has quit,it used to have millions of players.
This chart also shows how ladder and tournament play differ (quite greatly). VS ladder opponents, defensive macro play is harder because itâs reactive, meaning what you do depends on what your opponent is doing, and there is a wider variety of opponents on the ladder than on tournaments. That makes defensive macro harder. This is why Serral, a defense macro zerg, dominates tournaments but on the ladder zerg is literal trash. Zergâs defensive macro styles are too fragile vs all the possible opponents & each possible build each opponent could do. You canât achieve the same level of consistency because the opponent pool is larger.
The converse of this is that the race with the strongest defensive macro will be the strongest on ladder and the weakest in pro play, and thatâs obviously the case with protoss. A protoss makes a shield battery & a void ray, and all pre-lair zerg cheeses are nullified. Why? Zerg needs anti air to attack, and thatâs all locked behind Lair. Then everything at Lair-tech is countered by storm, again making it impossible to attack, which requires hive. Finally hive is achieved, allowing zerg to attack with ultras or broods, both invulnerable to storms. But this is equivalent to Protoss being invulnerable until hive by simply doing expand, void ray, expand, storm.
The statement is true and so is its converse. Clearly thatâs how the game works & it describes SC2 pro player & ladder play quite well, which gives it a unifying power that most theories from most people simply canât achieve. Itâs a good theory.
2 Likes
zerg is also the race that needs to play most actively for def.
harass toss once, next time there are 2xsb and 2-3xcannon.
now you have to commit to bane or more units.
in terran case, siege mode and forget it.
1 Like
Yep, zerg has to move units into position to defend. This is mainly because zerg is forced to expand faster than static defense is capable of defending. You are too spread out for static defense to be able to cover all the different attack vectors. You can build 5 spores per base and you will still take damage to a liberator, for example, so itâs just not possible to use static defense for defense. You have to use actual units, and that means active management of your defense with micro & positioning. By contrast, mutalisks hit a base a single time and the toss instantly warps in cannon/battery and the mutas are totally useless. At most you might be lucky to kill an assimilator or a straggler pylon.
100 gas (oracle) can kill an entire base of workers by simply focus firing drones but 1000 gas of zerg will be lucky to kill 3 probes on an assimilator using apm intensive stack-attack-stack micro.
2 Likes
They probably migrated to stormgate.
while zergs unit all have short range, you have to run into the range in T/P static def /unit if you want to do damage.
the commitment in Mutalisk does not feel good. If you lose, itâs usually because of muta â because they consume so many resources. Void/ BC/ oracle/ lib â all have the possibility to use defenses with impact. Muta does not.
If you win â you realize that only in very rare cases you win through muta alone.
btw: another example of zerg nerf, magic box has been removed.
2 Likes
Climate scientists are now accusing the industry of having âquasi religious beliefsâ and others are admitting that reaching a state of âneutralâ science is not only impossible but undesirable. Basically they are saying that everyone is biased so why pretend like we are not, and by the way our bias is a good thing. They try to split hairs between âscience free from valuesâ and âscience free from persona biasesâ but itâs a load of BS and thereâs no difference between the two. Where exactly do they think bias comes from if not their own values? This is exactly the kind of nonsensical religious style rhetoric that makes people think climate scientists are forming a new religion. They think that by changing the rhetoric used to refer to the same thing that it somehow makes that thing different, depending on the words used, and it doesnât. Bias is bias and it doesnât matter if you call it âpersonal valuesâ or anything else. Any part of the experiment that impacts the measurement of the dependent variable, other than the independent variables, is bias and it doesnât matter what you call it â itâs still bias. Bias is a category that includes your personal values as researchers.
The fact they are openly coming out and saying, more or less, âYeah, weâre biased, but thatâs a good thingâ should make everyone realize that this isnât a scientific process anymore â itâs a religious and political process.
This is why Iâve said in the past that the allocation of funding needs to be randomized. Right now researchers apply for grants and the grants are given out to fund research they think will be good for fundraising. Itâs obviously a political process. The way to fix this is that if a team of researchers apply for a grant, their request is added to a pool and when money becomes available a grant is given out at random. What this means is that if a climate denier scientist applies for a grant then he as equal odds of getting money as a climate affirmer. Thatâs whatâs necessary to remove the political bias.
There will still be some bias because scientists are a product of the culture they are raised in, so if they see climate affirming messages on social media all the time then thatâs obviously going to create bias. But if they fix the funding issue it will take care of 95% of the problem.
The reason itâs smart to give funding to climate deniers is because if climate change is as real as they claim it is then the deniers will come to the same conclusion. But they refuse to give them funding because they are scared it will produce a different result and that is religious style thinking. Itâs a purity test. The âinfidelsâ arenât given the privilege to participate in the âscientificâ process because they have the wrong beliefs. The funding process is a sacrament of a new religion and is not scientific anymore.