Miners are not moving around together as a group, their distribution and unpredictability make your method nearly impossible to implement in a reasonable amount of code. Please demonstrate how you'd accomplish this in a reasonable and uncomplicated way?
We'd simply insert a call to a function or a method right before the block reward amount to recognize for a particular block solution is finalized (this would be done by all the nodes), to check if the reward amount should be halved for this particular block solution reward or not. If not, the program executes the same as before. If it comes back yes, then divide the block reward by 2 but other than that, the program continues executing the same as before. Now, what would this function do? First, it would check if the extraordinary conditions exist (e.g. the current difficulty is more than 50% easier than the last difficulty). Extraordinary conditions would not come up too often, and if it hasn't, it would return false (difficulty fluctuation is normal, we don't need to engage the compensating mechanism). Now, if we detect the extraordinary circumstance, then we'd step back through the blockchain to count the number of block solution awards generated and credited to the same address during the last, higher difficulty phase (call it lastPeriodCount), and in the current, easier difficulty phase (currentPeriodCount). If lastPeriodCount < thisPeriodCount, then return true (cut reward in half). Otherwise, return false (don't cut in half).That's the full extent of the complexity of what I am proposing in terms of what the code would need to do at the coin network level - maybe 30-50 lines of code (if that). Pools would then factor this into how to distribute coins to miners. Showing that this mechanism is in operation in the pool and wallet UI is optional. I believe this is so powerful that it will become implemented by all the altcoins (and even Bitcoin itself) sooner or later, and will end the era of jumping coins minute-by-minute to try to maximize profits. It eliminates the need to keep changing around difficulty adjustment parameters to try to outfox the minute-by-minute coin hoppers - their paradigm is permanently defeated. After this change, long-term incentives would match up with short-term profit maximization. Both would agree that the correct strategy is to switch coins no more often than every few weeks or months, based on reassessing the various coins' long-term positioning and potentials, and the primary strategy would be to mine-and-hold, rather than mine-and-dump-on-the-exchanges. This is bringing back the original dynamic of Bitcoin miners, before there was any such thing as an altcoin. With this dynamic, market capitalization of cryptocurrencies can then truly "go to the moon," and we'd just be duplicating what was successful before.
I am happy with the fork. 
EU Stratum Server: stratum+tcp://eu.coinum.org:3336
US Stratum Server: stratum+tcp://us.coinum.org:3336