# In a game where a random success rate is offered (as opposed to a game that is 100% winable) what should you expect as a success rate?

Asked by rojo (24118) September 28th, 2018

On a solitaire app. there is a feature that keeps statistics. Four separate times now I have reached a success rate between 42% and 42.5% over 1000 games.

I have noticed that my success rate does increase over time and the number of games played. Would, at some number of games I reach a 50% success rate?

If you graphed it out with success over the number of times played what would the graph look like? Graph (note, the graph is for comparison purposes only. The data has no meaning in this discussion) or would it be reversed or a bell or what?

PS this is not homework. I have already completed that, including mopping the floors.

Observing members: 0 Composing members: 0

LOL! Interesting question. But you missed a spot on the floors.

Well, you have to factor in that it’s not just random success. There is some strategy involved too.

Dutchess_III (40373)

I see a HP graph for 2.0 TSI engine ? ? ?

By the way my Solitaire success rate is 43% !

@Tropical_Willie I tried to say that the information contained in the graph is of no consequence to the question; I am only interested in how the data is shown. My question with the graph is whether the distribution of data points of the percent success rate over the number of games played would be of a similar nature where you start out with either a 100% success rate or a 0% after your first game depending upon whether you won and that any successive games would only increase/decrease the percentage rate of success depending upon your starting point. I suppose it could either be this way or the opposite depending on whether you were starting from 0 or 100 but either way the average would flatten out the curve to the same point; regardless of starting point. Is that correct?

rojo (24118)

@rojo, that i correct, and your success rate will generally revert the mean

To increase your overall success rate to 50%, you would have to play another 1,000 games at a success rate of 57.5% to 58%. That would be a huge improvement from your current rate.

zenvelo (33773)

@zenvelo how would you determine the mean in a problem like this?

rojo (24118)

@rojo the mean is 42.5%, that is your success rate over 1,000 games. So unless you suddenly get some insight on playing the game that bolsters your victory rate, you will tend to average 42.5%.

zenvelo (33773)

Some people manage an 80% win rate.
I read somewhere, that the percentage of solvable games is between 82% and 92%.
So really, your 42% percent win rate is a reflection of your own skill level.

ragingloli (45308)

^^ that is what I said.

Dutchess_III (40373)

You asked the question as if it would have a general answer for different games that have a random success rate. Different sorts of games would have both different percentages possible, and different learning curves. Different people would also show different learning curves for the same game, It’d also vary by things like how tired they are when playing and how much attention they are paying, and whether they are trying different techniques.

Zaku (23131)

I bet those long winter nights just fly by…

ucme (47198)