My reminder on entropy

It is defined that every event seems equally surprising, or unexpected.  Nothing seems repeatable or regular.

We see this chaos when the system has already been 'ordered' ordered in the sense of all redundancy is removed.  If the system were described in its index space, then infrequent large events take up more space, leaving less for the frequent small.  But at maximum entropy, each index event is seems independent.

To maximize entropy is to reduce redundancy and release the chaotic residual. Like in the NBA stats.  The bet was uniform random pick within groups, but between groups the odds were large, and quantized.  Thus, on net, the odds weighed out the bet size and all bets, over all, seem uniform random picks.

Best to bet the whole channel, wee are back to the debate in maximum entropy betting.  Looking at he published NBA statistics we can see the quantization of the queue intt three tiers.  One wants to bet each tier equally, using payoffs according to tier order.  Really, the published NBA percentages are already ordered by owners doing trades, take their bet at face value. This is like portfolio rebalancing, the portfolio contains bets on each of the asset classes, treated as semi-independent components in a confined channel. If you can observed the structured buy and sell queues, you can compare queues by ratio. Stable queues minimally redundant, maximum entropy.