That’s an availability bias, a cognitive bias that occurs when people judge the likelihood of an event based on how easily examples come to mind.
Basically if someone finds something rare: they will talk about it and make a thread about it.
It’s literally a statistic that was presented. There is no bias. It’s from the same hsgurureplay or whatever that you hang your own hat on.
1 Like
You don’t say anything new. The word statistic doesn’t make the sample it has large enough to have a low error.
18 samples have extremely high errors in this game.
Take a look at decks with lower than ~60 samples,
the win rate jumps all over the place.
I don’t think you get it. The statistic is already far and away proving the point that was presented;aka that painter virtue always shows up by turn 4 even in the absence of tutoring. Even IF we extrapolated the statistic to 30, and even IF we assigned it the opposite/worst we could (aka painter virtue basically never showing up or showing up too late)…it would still be an 18-12 statistic which is still 60% that they would have it by turn 4. If we took a more moderate path of extrapolation, then we’re at 24-6, or 80%. And of course the most extreme path in favor, leaves us at 30-0, 100%. The point has already been made and still stands even in the worst case scenario. Paladin always has painter virtue by turn 4.
1 Like
What are you talking about. You lack basic fundamental knowledge on probability here. For example you may have 1% chance for what happened to happen and it has happened.
It doesn’t mean it doesn’t happen 99% of the time otherwise.
I can give you mathematical examples if you want.
Idk what you are talking about. I understand all of that. I have a master’s degree in mathematics.
1 Like
Just an update, I have since my last post played another 7 Paladins. Again, 100% of them got this card in their opening hand. One can claim its not a good statistical sample but this many with a deviation of ZERO clearly is a problem. Any class with a guaranteed card in their opening hand is broken and requires zero skill. Of course it is possible to counter it, if you get the right draws. If not, the game is decided before it even starts.
2 Likes
If you want to learn Statistics and Probability go study them. Don’t guess.
No: 7 is not a sample that anyone respects even for a simple coin toss.
And this isn’t a simple coin toss. The population is extremely skewed.
18+7=25 ±5 Standard Deviation = their data is valid
1 Like
For highly skewed populations even 25 samples may not be representative. Skewness means the majority of the outcomes might cluster in one region and rare events won’t show up until the sample size is much larger. That’s why small sample sizes can easily mislead, even if the data shows no apparent deviation.
Imagine flipping a biased coin with a 95% chance of heads and 5% chance of tails. Even after 25 flips: you might see heads nearly every time and it would look like tails never occur. It’s only with hundreds of flips that you’d start seeing the true proportions emerge and that’s why 25 is not enough for skewed distributions like this.
PS You mentioned a standard deviation of 5. Was this calculated based on the data or was it an assumption? I ask because it’s tricky to estimate a meaningful standard deviation with such a small sample size, especially for a skewed population.
My point was you were wrong going after them for the 7 sample when they said since the last post they played another 7, well their last post they played 18, that gives us 25, not 7. Seeing as 25 is close to 30 (±5), and 30 is often used as a rule of thumb for a minimum sample size needed to achieve statistical significance, my point that you were wrong in going after them is true/valid
1 Like
The problem isn’t the sample size, it’s whether the data was fairly obtained (i.e. whether that truly is all of the games they played against paladin). If they didn’t, and there’s other games being omitted, then there’s confirmation bias which skews the data.
On the other hand, if it is then at 25/25 games in which the paladin drew the weapon (by turn 4 I assume is what’s being claimed), then that’s a 93.54% credence that the true probability of getting the weapon by turn 4 is between 90-100%. Although with them having tutors and drawing 7 cards (plus mulligans), that wouldn’t actually be particularly surprising.
1 Like
18 times in a row is an unlikely occurrence for any individual player. But when a quarter million people play the game each day, it becomes more likely that such an unlikely occurrence can happen to someone.
I’m sure you know this since you have a degree in mathematics.
1 Like
25 times* and yes that’s basically at the bar for where we look at it as statistically sound and relevant
1 Like
I’m not sure you understood what they told you. Reread it. They’re telling you something being unlikely isn’t impossible.
By definition being unlikely is not impossible. If that was not true it would not be called unlikely but impossible.
Unfortunately (for you) I did read it and understand it, as I also have a degree in English; though I reread it just because you asked, and learned nothing new. Unlikely things to the point of nearly impossible happen every day. Maybe you need to reread what I’ve written and realize what you’re even talking about before saying anything more.
Then rewrite how you responded, because they are talking to you about likelihood, and you are talking about efficacy of a specific sample size.
I…don’t care what they are talking about. I’m talking about what I’m talking about. They are the ones who need to rewrite how they respond, not me (if anything)…then you come in here wasting my time, per usual. I guess I should have known by now, that’s on me
1 Like
And btw you’re still wrong about efficacy. This isn’t a simple coin toss. You’ve yet to study how sample size efficacy is affected by skewness.
Or I’m not wrong, and perfectly aware. But congrats to continuing to waste my time, you’re quite skilled