So on Sharkscope you can find out the average Skill Level for a Tourney, and I was wondering how good you have to be to be a winning player on a given field.
So if the avg skill level of the players is 70. You probably wont make money on if you are just 72. I tried to find a way to determine how much better than the average player you have to be. Therefore I had to do some discussable assumptions:
-- You have to be one of the best 10% of players to profit on a tournament
-- The skill level is distributed normal (If you draw a random player you have the same probability hes 5 levels lower than average as higher than the average)
-- I assume that the range is distributed in between a range thats determined by the lower difference to either 50 or 99. So with 70 the difference to 50 is 20 and to 99 its 29. So I assume the 3 deviation range is from 50-90.
That are the best assumption I could made. Im happy if someone can make better one.
So with that figures i just divide the range limit by 3 and have a standard deviation. (20/3 = 6.6666..7). If I now want to find the Skill Level to be one of the TOP 10% players you just calc AVG Skill Level + 1.281552 * Standard Deviation (yeah that's the value inside which 80% of all values in a normal distribution are). And we would end up with 78,54 here.
I also made a small jsfiddle so you just have to type in the avg skill to calculate the required skill.
How to improve the formula
So to make the formula more accurate the most important steps would be
a) Gather Information about the standard deviation of skill levels in a tourney.
The only way to do it currenty, would just like click either all players of one or more tourneys and gather data points paired with the avg skill of the tourney or if possible click random players and gather data points. Its an enormous task Im currently not willing to do lawl. Maybe a crawler could do this or I talk to Sharkscope, maybe they can build a new feature or stuff.
b)Determine the % of winning players in a tourney
To determine this exactly you would probably need even more data than for a). I thought maybe a good solution would be to find the lowest rank player on officialpokerrankings that is still a profitable player. And just check his percentage. That would be a pretty educated guess. Maybe some more experienced player can guess better than 10% here. It should also be different for different structures. Since more edge is possible with bigger stacks and more time.
-- be aware that tourneys decrease in average skill the longer they run. Mostly regs reg before a tourney start. So if you check a not yet started tourney on Sharkscope, better take the same tourney from yesterday.
-- different times different skill levels. In the evening the fields are a bit tougher than in the night and in the morning.
Im happy over every point you find to critizise. See this just as a first sketch.