| Simulate |
Start a simulated run (a session of plays). |
| Ante |
Ante is the cost of one play. It may very well be 0, which means that
it is just the probabilities and the reward values which is used.
|
| Values |
This is a "," (comma) separated string for the reward/losses of the game.
E.g. the values
-1,1
is the values for simple coin toss, where (say) head gives a loss of 1
and tail gives a gain of 1. The value vector of
-1,-1,1,1,1,1
may be seen as the game where you win 1 if a dice thrown shows 3,4,5,6 and you
loose 1 if it shows 1 or 2.
|
| Probabilities |
The probability vector is a "," (comma) separated string, e.g.
0.5,0.5
which
is the even probabilities for (e.g.) a coin toss. If the numbers don't add up
to exactly 1 (as valid probabilities must), the program scales (normalizes)
the values to proper probabilities. E.g.
1,1
is the same as the probabilities
0.5,0.5
A probability vector of
1,2
assigns the first value twice chances of the second value over the first,
i.e. is the same as the probability vector
0.33,0.66
|
| Statistics |
The statistics is:
- Acculated gain: The final score. This is calculated for each play
as such:
- The player always starts at 0.
- For the given (randomized) value get the reward (from the values field)
- Add the values of reward - ante to the existing gain.
- Expected (gain): This is the theoretical gain, calculated
by first get the mean value of the rewards given the probability for each
reward, and then
multiply this value with number of runs.
- Min value: The minimum value in this run
- Max value: The maximum value in this run
- Num zeros: The number of zeros is the session.
- Times in lead: Number of times the player was in the lead.
If the accumulated score in the run is 0, we count it as lead only if the
player was in the lead (i.e. was above 0) the former time.
|