Hidden Markov Models

The Gist

A HMM is a random process that emits events when triggered and also encodes the probability of a transition to another HMM. An observer cannot tell which is the HMM in effect, hence its hidden nature. In statistical modeling, HMM are used to reconstruct the sequence of hidden states based on a sequence of observations (emissions). In this mod, however, there is no need to infer hidden states so they are used as a simulation tool.

Balancing

Balancing is done using matrices as typically done to find steady state probabilities of a state in a HMM (See HMM balancing .maple file). From the steady state probabilities, the expected balance of emission is computed and projected over 10 years of random walk.

space_craze

Currently balanced to be neutral for attention_span with and 10% of one increment of loss over 10 years. This property should always be balanced and the base period time should be used to refine instead. As for hype, the odds are against the players with an expected 5 increment loss over ten years. This makes sense since a program not accomplishing anything for long while deserves to be underrated.