# Simulate data for both groups
group_1 <- rgeom(10000, 1/(1 + 3.7))
group_1_alternate <- rgeom(10000, 1/(1 + 3.7))
group_2 <- rgeom(10000, 1/(1 + 4.35))

# Compare Group 1 to itself to get a baseline chance that a random person
# drawn from this distribution outperforms itself
group_1_greater_self <- sum(group_1 > group_1_alternate)

# Compare Group 2 to Group 1 to get the new percentage chance that a random
# person drawn from Group 2 outperforms a random person from Group 1
group_2_greater_group_1 <- sum(group_2 > group_1)

# Calculate the risk ratio
training_rr <- group_2_greater_group_1/group_1_greater_self

# The risk ratio falls around 1.09. Can I interpret this that there's a 9%
# chance someone going through training will see greater sales than if they
# didn't attend training?