Stop crying about russian bias

[Top Goy]

-Most of these "arguments" come from Thunder Skill. The stats Thunder Skill provides are not reliable. (why in the video linked)
-" No spalling APFSDS against russian armour" congrats dude that happens with russian APFSDS aswell. Its even worse for Russia since western tanks are spaced
-"Russian armour eats shells (example: T-34 hatch)" congrats dude, you discovered the fact that volumetric shells have bugs
- "muh KA-50 spam" spawn in a plane
-"cant spawn in a plane cuz of pantsir" ever heard of going behind a mountain? also i love how they pretend that ADATS or other AA cant snipe russian airplanes out of the sky

In general, people like these dont know how to play the game, they get their stats from an unreliable source and they forget that the problems they adress (for example shells being eaten, apfsds not spalling) apply to any other nation the same. I play both germany and russia and i can tell for sure that i had these problems on both nations.

anyway, there is no russian bias, please go outside:

https://www.youtube.com/watch?v=5lBIbxQmg4Y

[Velico]

Nice. I'll just copy pasta my response to you about this ten minutes ago since you had nothing to refute.

I've watched this video and commented on its inaccuracies. The fact the video was made at all is a farce. EC makes a living off War Thunder content and is part of their partnership program. His potential conflict of interest is very high, especially with the verbiage within the video. He is amplifying Gaijin's attempt to undermine credibility while offering no information themselves to be transparent and/or accountable. So I firmly reject most of what EC has to say in this video, as his rhetoric is very much in line with anti-intellectualism.

Their argument that people would not be able to comprehend statistics is rooted in anti-intellectualism and is unfounded; it is PR speak, and you are amplifying it with this link. EC saying a dataset is "useless" and "faulty datasets are way more damaging than no datasets at all" is completely off-base. Data is data, and all of it is useful in an increasingly data-driven world. Those who weaponize data and draw erroneous conclusions are troublesome, but that does not need to be the only goal of statistics. In this case, we can use it to pressure Gaijin into being more accountable and transparent, which is the trend in gaming, especially with titles vying for eSports.

Thank you for the extra exposure into Gaijin's failure to the playerbase.

[Velico]

Originally posted by черный русский:

Originally posted by Velico:

Nice. I'll just copy pasta my response to you about this ten minutes ago since you had nothing to refute.

I've watched this video and commented on its inaccuracies. The fact the video was made at all is a farce. EC makes a living off War Thunder content and is part of their partnership program. His potential conflict of interest is very high, especially with the verbiage within the video. He is amplifying Gaijin's attempt to undermine credibility while offering no information themselves to be transparent and/or accountable. So I firmly reject most of what EC has to say in this video, as his rhetoric is very much in line with anti-intellectualism.

Their argument that people would not be able to comprehend statistics is rooted in anti-intellectualism and is unfounded; it is PR speak, and you are amplifying it with this link. EC saying a dataset is "useless" and "faulty datasets are way more damaging than no datasets at all" is completely off-base. Data is data, and all of it is useful in an increasingly data-driven world. Those who weaponize data and draw erroneous conclusions are troublesome, but that does not need to be the only goal of statistics. In this case, we can use it to pressure Gaijin into being more accountable and transparent, which is the trend in gaming, especially with titles vying for eSports.

Thank you for the extra exposure into Gaijin's failure to the playerbase.

funny part is i had something to refute
you kinds of people really need to get a life outside, and maybe just maybe realize that there is no russian bias
I can also go use false fcking statistics to say that the US has an 80% winrate

American Bias everyone:lunar2019grinningpig:

For the record, I never said anything about Russian bias, you assumed that. I merely pointed out that any vehicle that is below 46% or above 54% WR is known as a statistical outlier. And beyond those are extreme outliers. When that happens, changes need to be made for greater parity, but Gaijin refuses to do this. Go find another dataset of WT data that has your example and we can go from there. Amusing that you tell others to get a life after defending Gaijin with utmost vigor.

[Top Goy]

Originally posted by Velico:

Originally posted by черный русский:

funny part is i had something to refute
you kinds of people really need to get a life outside, and maybe just maybe realize that there is no russian bias
I can also go use false fcking statistics to say that the US has an 80% winrate

American Bias everyone

For the record, I never said anything about Russian bias, you assumed that. I merely pointed out that any vehicle that is below 46% or above 54% WR is known as a statistical outlier. And beyond those are extreme outliers. When that happens, changes need to be made for greater parity, but Gaijin refuses to do this. Go find another dataset of WT data that has your example and we can go from there. Amusing that you tell others to get a life after defending Gaijin with utmost vigor.

"you never said there is russian bias" yet you for some reason to dispute the stupid claim of a 70% winrate from unreliable data.
I know gaijin is bad at balancing ♥♥♥♥, and you might be right about your 46% or above 54% winrate. gaijin does a lot of sht wrong, but to assume that they prefer russia over any other nation is stupid and childish. people who unironically believe that cant be further from wrong. And the data of russia winning 70% of all games is plain old wrong.

[DAWHIPPA]

Downlaoding again after taking about 2 years off. I see the Russian bias thing is still going on. lol. I do remember the Is-6 being under tierd at 6.7 and the KV-1E OP. What other tanks are OP now? The top tier ones were always tough from the front but all of them could get 1 shotted very easily from the ammo Carousel

[Velico]

Originally posted by черный русский:

Originally posted by Velico:

For the record, I never said anything about Russian bias, you assumed that. I merely pointed out that any vehicle that is below 46% or above 54% WR is known as a statistical outlier. And beyond those are extreme outliers. When that happens, changes need to be made for greater parity, but Gaijin refuses to do this. Go find another dataset of WT data that has your example and we can go from there. Amusing that you tell others to get a life after defending Gaijin with utmost vigor.

"you never said there is russian bias" yet you for some reason to dispute the stupid claim of a 70% winrate from unreliable data.
I know gaijin is bad at balancing ♥♥♥♥, and you might be right about your 46% or above 54% winrate. gaijin does a lot of sht wrong, but to assume that they prefer russia over any other nation is stupid and childish. people who unironically believe that cant be further from wrong. And the data of russia winning 70% of all games is plain old wrong.

Ok, clarification since you're showing sincere effort. The reason the data from Thunderskill is viable to draw from is two-fold: 1) the sample size is 3,714,218 players in thunderskills database, which is a statistically significant amount of people and Gaijin does not provide any statistics on this. And 2) the amount of new users registered and statistics are continually updated. I made a comment 3 months ago with this number as reported from thunderskill: 3,391,124. Compared to the above, that is 330,245 new users data collected in that time, or about 100k per month.

A margin of error tells you how many percentage points your results will differ from the real population value. For example, a 95% confidence interval with a 4 percent margin of error means that your statistic will be within 4 percentage points of the real population value 95% of the time. So when any (any*) vehicle is less than 46% or greater than 54% it is known as an outlier. In simple terms, an outlier is an extremely high or extremely low data point relative to the nearest data point and the rest of the neighboring co-existing values in a data graph or dataset you're working with.

While there are some issues with quantitative coding (used for analysis), Gaijin offers no transparency with their statistics and this is the best available information out there to draw from. Data is data, and all of it is useful in an increasingly data-driven world. I (and others) are simply pointing out that statistically this shouldn't happen, and that it needs to be addressed. How it is addressed is a different matter.

[Top Goy]

Originally posted by Velico:

Originally posted by черный русский:

"you never said there is russian bias" yet you for some reason to dispute the stupid claim of a 70% winrate from unreliable data.
I know gaijin is bad at balancing ♥♥♥♥, and you might be right about your 46% or above 54% winrate. gaijin does a lot of sht wrong, but to assume that they prefer russia over any other nation is stupid and childish. people who unironically believe that cant be further from wrong. And the data of russia winning 70% of all games is plain old wrong.

Ok, clarification since you're showing sincere effort. The reason the data from Thunderskill is viable to draw from is two-fold: 1) the sample size is 3,714,218 players in thunderskills database, which is a statistically significant amount of people and Gaijin does not provide any statistics on this. And 2) the amount of new users registered and statistics are continually updated. I made a comment 3 months ago with this number as reported from thunderskill: 3,391,124. Compared to the above, that is 330,245 new users data collected in that time, or about 100k per month.

A margin of error tells you how many percentage points your results will differ from the real population value. For example, a 95% confidence interval with a 4 percent margin of error means that your statistic will be within 4 percentage points of the real population value 95% of the time. So when any (any*) vehicle is less than 46% or greater than 54% it is known as an outlier. In simple terms, an outlier is an extremely high or extremely low data point relative to the nearest data point and the rest of the neighboring co-existing values in a data graph or dataset you're working with.

While there are some issues with quantitative coding (used for analysis), Gaijin offers no transparency with their statistics and this is the best available information out there to draw from. Data is data, and all of it is useful in an increasingly data-driven world. I (and others) are simply pointing out that statistically this shouldn't happen, and that it needs to be addressed. How it is addressed is a different matter.

Its so funny that you say data is data but dont even think for a second that that data might be false. Really funny choosing to ignore all of the points mentioned by that vid, and how the way these "statistics" are made are flawed from the beginning on. God im so ♥♥♥♥♥♥♥ tired of this stuff, this entire "argument" of russian bias is such an old meme at this point. I agree that this game doesent balance stuff very well, but i actually feel sad for clown that unironically think that russian bias is an actuall thing. Im done responding its actually killing me inside already

[Velico]

Originally posted by черный русский:

Originally posted by Velico:

Ok, clarification since you're showing sincere effort. The reason the data from Thunderskill is viable to draw from is two-fold: 1) the sample size is 3,714,218 players in thunderskills database, which is a statistically significant amount of people and Gaijin does not provide any statistics on this. And 2) the amount of new users registered and statistics are continually updated. I made a comment 3 months ago with this number as reported from thunderskill: 3,391,124. Compared to the above, that is 330,245 new users data collected in that time, or about 100k per month.

A margin of error tells you how many percentage points your results will differ from the real population value. For example, a 95% confidence interval with a 4 percent margin of error means that your statistic will be within 4 percentage points of the real population value 95% of the time. So when any (any*) vehicle is less than 46% or greater than 54% it is known as an outlier. In simple terms, an outlier is an extremely high or extremely low data point relative to the nearest data point and the rest of the neighboring co-existing values in a data graph or dataset you're working with.

While there are some issues with quantitative coding (used for analysis), Gaijin offers no transparency with their statistics and this is the best available information out there to draw from. Data is data, and all of it is useful in an increasingly data-driven world. I (and others) are simply pointing out that statistically this shouldn't happen, and that it needs to be addressed. How it is addressed is a different matter.

Its so funny that you say data is data but dont even think for a second that that data might be false. Really funny choosing to ignore all of the points mentioned by that vid, and how the way these "statistics" are made are flawed from the beginning on. God im so ♥♥♥♥♥♥♥ tired of this stuff, this entire "argument" of russian bias is such an old meme at this point. I agree that this game doesent balance stuff very well, but i actually feel sad for clown that unironically think that russian bias is an actuall thing. Im done responding its actually killing me inside already

I addressed the main points in refutation above. You are now acting in bad faith and I am done trying to educate you.

[+rep Vollidiot]

Originally posted by черный русский:

-Most of these "arguments" come from Thunder Skill. The stats Thunder Skill provides are not reliable. (why in the video linked)
-" No spalling APFSDS against russian armour" congrats dude that happens with russian APFSDS aswell. Its even worse for Russia since western tanks are spaced
-"Russian armour eats shells (example: T-24 hatch)" congrats dude, you discovered the fact that volumetric shells have bugs
- "muh KA-50 spam" spawn in a plane
-"cant spawn in a plane cuz of pantsir" ever heard of going behind a mountain? also i love how they pretend that ADATS or other AA cant snipe russian airplanes out of the sky

In general, people like these dont know how to play the game, they get their stats from an unreliable source and they forget that the problems they adress (for example shells being eaten, apfsds not spalling) apply to any other nation the same. I play both germany and russia and i can tell for sure that i had these problems on both nations.

anyway, there is no russian bias, please go outside:

https://www.youtube.com/watch?v=5lBIbxQmg4Y

keep up with the bs
adats can get the sams just up to 65 degres? because of war thunders dumb ideas
Next russian apfsds deal the highest dmg
Let me guess bigger gun bigger apfsds rod right ? nope

and funny how only russia has those "whoopsis" everytime
keep up with your fantasy version of stronk russia.

[Crewmember 123]

Originally posted by черный русский:

Its so funny that you say data is data but dont even think for a second that that data might be false. Really funny choosing to ignore all of the points mentioned by that vid, and how the way these "statistics" are made are flawed from the beginning on. God im so ♥♥♥♥♥♥♥ tired of this stuff, this entire "argument" of russian bias is such an old meme at this point. I agree that this game doesent balance stuff very well, but i actually feel sad for clown that unironically think that russian bias is an actuall thing. Im done responding its actually killing me inside already

One guy is providing viable arguments why thunderskill is a reliable source. the other one is you.

[torindechoza]

Originally posted by Velico:

Ok, clarification since you're showing sincere effort. The reason the data from Thunderskill is viable to draw from is two-fold: 1) the sample size is 3,714,218 players in thunderskills database, which is a statistically significant amount of people and Gaijin does not provide any statistics on this.

You said that you refuted the video. But:
15:42 in the video TEC explains why you cant extrapolate.
and earlier he explains how bad players do not look their own stats there, therefore their stats are not counted.

im looking at thunderskill
The average player in realistic has:
1.6 K/D
53.78 Winrate
Who is getting killed there? Who is losing?
those are obviously false statistics.

edit:
*its funny that players want the devs of the game make balance changes based on incomplete data, when they have the complete data.
*some players will complain about the T34 driver hatch but not about the tiger´s edge of side armor that will eat your shell, or the leopard's fron armor overperforming.

at the end the game has bugs and balance issues but the majority of complaints are ridiculous.

[shipdesignerdude]

Originally posted by torindechoza:

Originally posted by Velico:

Ok, clarification since you're showing sincere effort. The reason the data from Thunderskill is viable to draw from is two-fold: 1) the sample size is 3,714,218 players in thunderskills database, which is a statistically significant amount of people and Gaijin does not provide any statistics on this.

You said that you refuted the video. But:
15:42 in the video TEC explains why you cant extrapolate.
and earlier he explains how bad players do not look their own stats there, therefore their stats are not counted.

im looking at thunderskill
The average player in realistic has:
1.6 K/D
53.78 Winrate
Who is getting killed there? Who is losing?
those are obviously false statistics.

edit:
*its funny that players want the devs of the game make balance changes based on incomplete data, when they have the complete data.
*some players will complain about the T34 driver hatch but not about the tiger´s edge of side armor that will eat your shell, or the leopard's fron armor overperforming.

at the end the game has bugs and balance issues but the majority of complaints are ridiculous.

People just choose data when it only suites their argument.

Why can't everyone just agree that most things in this game are buggy and that everyone gets affected by it?

[Velico]

Originally posted by torindechoza:

Originally posted by Velico:

Ok, clarification since you're showing sincere effort. The reason the data from Thunderskill is viable to draw from is two-fold: 1) the sample size is 3,714,218 players in thunderskills database, which is a statistically significant amount of people and Gaijin does not provide any statistics on this.

You said that you refuted the video. But:
15:42 in the video TEC explains why you cant extrapolate.
and earlier he explains how bad players do not look their own stats there, therefore their stats are not counted.

im looking at thunderskill
The average player in realistic has:
1.6 K/D
53.78 Winrate
Who is getting killed there? Who is losing?
those are obviously false statistics.

edit:
*its funny that players want the devs of the game make balance changes based on incomplete data, when they have the complete data.
*some players will complain about the T34 driver hatch but not about the tiger´s edge of side armor that will eat your shell, or the leopard's fron armor overperforming.

at the end the game has bugs and balance issues but the majority of complaints are ridiculous.

Extrapolation - TEC is incorrect in this because the sample size, which he does not disclose either out of ignorance or malice, is well beyond what is necessary to tabulate and analyze.

While the average player on thunderskill may be above average, it is still within the average range, meaning low deviance. It is likely that with more data those numbers would fall closer to 50/50 parity, but we do not have access to that data because Gaijin chooses not to disclose it.

That does not make those statistics "obviously false", it makes the data reliable but incomplete. Incomplete data is still valuable for understanding trends over time, and this is the point.

[torindechoza]

Originally posted by Velico:

While the average player on thunderskill may be above average, it is still within the average range, meaning low deviance.

"within the average range"
1.6 K/D is 60 % more than average.

edit: i feel very good when i see my thunderskill that shows me im a fantastic player.
but then i see the actual statistics that the game gives me (in the player card) and im a little less good there.
Then i can understand why people like the site so much. Its an ego booster.

[Velico]

Originally posted by torindechoza:

Originally posted by Velico:

While the average player on thunderskill may be above average, it is still within the average range, meaning low deviance.

"within the average range"
1.6 K/D is 60 % more than average.

Not by much, but it is an outlier. WR is a better statistic to follow for evaluating than KDR for overall performance of vehicles and game mechanics. So let's create an example.

Thunderskill represents ~40% of the total playerbase, and has more above average players in its database. While that dataset may hold inflated numbers to what would be expected, it does not nullify that it represents a portion of the playerbase. All it means is that it's on the high end of the law of averages due to a variety of variables I do not have access to. Repeated trials will converge to a stable value over time as the number of samples increases, so if that 40% were made 100%, the increase in samples taken would bring those numbers down. This is the Law of Large Numbers. Bernoulli proved in a precise manner by showing that, as the number of repetitions increases indefinitely, the probability of this fraction being within any prespecified distance from p (probability of a win) approaches 1.

EDIT: Corrections and additions.

[Velico]

Originally posted by torindechoza:

edit: i feel very good when i see my thunderskill that shows me im a fantastic player.
but then i see the actual statistics that the game gives me (in the player card) and im a little less good there.
Then i can understand why people like the site so much. Its an ego booster.

Interesting, can you link your queries? Mine are the same.

https://warthunder.com/en/community/userinfo/?nick=Velico85

https://thunderskill.com/en/stat/Velico85