Using this program, you can simulate an outcome of matches in the long run given several starting variables. You can download exe file from releases or you can compile the source on your own.
When you run the program:
Enter the number of games to simulate (1..2147483647), 0 to quit: I recommend somewhere around 1000000 tries, but it depends on the speed of your computer.
Enter the probability of Team A winning rounds with their skill (0..1): You probably want to start with equally strong teams, so 0.5 should be good. Be careful that this is the probability of team winning a single round, so if you have anything other than 0.5, the resulting probability of winning an entire match is much more pronnounced.
Enter the probability of team winning rounds on the side Team A started given the map imbalance (0..1): Some maps in CS are not balanced. For example if you want to simulate a game on a de_nuke, you might want to enter something close to 0.7 (or 0.3 - depending on which side you want your team to start).
Enter the probability of Team A winning pistol rounds (0..1): The same as pevious, but only applies to pistol rounds. A map can be more balanced in pistol rounds than in other rounds.
Enter the probability of a team winning the next round given they won the previous round (momentum) (0..1): This simulates momentum. If a team is in a momentum it is likely to win the next round if they won the previous round. I may be simmilar to autocorrelation. Setting it to 0.5 is like there is no momentum in the game, setting it to closer 1 is like a positive autocorrelation (a team is more likely to win ne next round if they won the previous one) and setting it to number closer to 0 means the opposite.