Visualization of epsilon greedy on a 3 armed multi armed bandit.
The sliders control the payout per bandit 0-100, bandit size adjusts accordingly.
Epsilon controls ratio of exploration vs exploitation 0-1.
Bandit color indicates fraction of times this bandit was chosen over the last 20 pulls.
Click "Run" to start the simulation, "Stop" to pause. Reload page to reset.
Source code on Github.