Please provide a method of increasing sample size.
PS: don’t forget that there is a limited number of pros.
1 Like
The only way is to either look at other major tournaments. Or if you somehow had the data, look at ladder from Dia 1 to GM.
1 Like
This what you suggested but OK let’s I will fallow. Round of 32 is a group match. The group consists of 4 players, each player plays 3 matches in each match are 3 rounds.
_2019 WCS Fall - Liquipedia - The StarCraft II Encyclopedia
Total 3 * 3 = 9.
You have around 1000 pros base on aligulac (all players that are recorded). So how do you plan to increase sample size?
OK so what is number 400
Why you do the difference of skill level “Rb - Ra”
Why you do 10 to power
And what is 1 +
I don’t disagree with Math.
I disagree with you. Its a BIG difference.
No, I do not deny the validity of the probability space. What I deny Is your incorrect use of probability space.
- Zerg, Terran, Protoss are imaginary creations.
2.) The player selects race. - SC2 tournament is an individual competition consisting of 1v1 matches.
- player that defeat all his opponents win
Example: If Serral defeats all his opponents in the tournament and win Then the race Serral played also win.
No, it’s not.
The program gives probabilities of each team winning each round in the tournament.
This would include the probability of winning the championship for each team.
The probability of team i winning through is the probability that i won through round r -1 times the probability that i beats its opponent in round r.
In round one there is one possible opponent, but the number of possible opponents increases in later rounds,
up to half the total number of slots in the championship round. Therefore,
the opponent of team i is conditional on winning the previous rounds. In general, for a single elimination tournament
But I leve it at that. Everyone have access to the paper and can check it.
TAKE NOTES NOOBS! This is how you torll! these newbs are still fighting about this!!! LMAOOOOOOOOOOOOO
Unless the player is not a pro he was included in 1000 Pros.
If you include Ladder then you need to prove that ladder meets your Criteria. if you do not prove that you make your data set questionable those conclusions invalid.
Lmfao this man speaks the truth
Like everyone else, you’ve taken this a step too far.
The basics of statistical analysis is having an adequate sample size. You do not have an adequate sample size with just pro tournaments.
I’d be shocked if Blizzard didn’t use ladder statistics at all.
I didn’t say you are not allowed to use non-pros.
What I say is that the population can be added to your sample but you need to ensure that this population meets the criteria of the initial sample.
If you are unable to find additional data to add to your sample size, then:
A) you have an appropriate sample size
B) your criteria and/or purpose are incorrectly
As blizz has demonstrated on their twitter account, they have some kind of data collector coded in for the purpose of collecting win rate percentages. It would be strange to have that but decide not to use it.
Regarding sample size, I think it doesn’t even matter. If you have data on the entire population, samples aren’t taken. Population too small? That doesn’t make sense. If a population of 1 existed, that 1 person would be definitively representative of himself because he is all there is. If 500 pros exist, they will be definitively representative of themselves. When you have a sample, you run the risk of it not being representative, so you would have to worry about sample size.
So, yes, you are correct. Despite that, you still don’t need to worry about having an adequate size since you have the whole population.
1 Like