.
.
.
1. Do we trust the data (charts on blockchain.info and other places) and assume this wide fluctuation in network capacity is real?
Charts data on blockchain.info is based on number of blocks found, not on the hashrate of the hardware (or pools)..
.
1. Do we trust the data (charts on blockchain.info and other places) and assume this wide fluctuation in network capacity is real?
Very lucky period for some big(er) pool, or few of them, can influence charts very much.
I tracked Eligius for few days at end of March: with realtime hashrate reported on stats page Eligius should mine 22 blocks per day, and there were days with 27, 29, 30 blocks.
When such string of lucky days happens on some even bigger pool, we can see that wide fluctuation.
Green line on the chart at https://bitcoinwisdom.com/bitcoin/difficulty doesn't have any wide fluctuations - that's charted hashrate based on 2016 blocks found - any wide fluctuaction averages out.
Thanks for the clarification. Did some more research and you're right about the chart data. Was hoping there was more data available, but the network hash rate is mostly derived from the only metric that is public: Avg. time between blocks mined, which can vary quite a bit on a daily basis.