Given that the hashrate graphs you see online are all estimated based on difficultly and the rate of blocks being mined, these graphs misinterpreted this gap as a drop in hashrate, when in actual fact it was a perfectly normal delay in blocks being mined.
well, not quite. the graphs were calculated perfectly well
If you look
only at a very short period hashrate graph (say, 24 hours of data), then you only have very short windows over which to calculate the averages (e.g. a 24 hour period might use 1 line for the average over 4 hours, a 2nd line for the average over 2 hours and then a 3rd line for the average over 1 hour).
The misinterpretation was human, and deliberately so to exaggerate the effects of only 1 long block, 1 long block amongst many shorter than average blocks. If you focus on the shortest possible window to calculate the average predicted hashrate, of course it's going to indicate more extreme swings in variability, an average with very few data points is inherently more sensitive to higher variance than averages with larger amounts of data points.
It reminds me a lot of people who get panicked when they see a transaction of 10,000 bitcoins and are convinced the market is about to dump, and every time it turns out to be a complete non-story and just some exchange just moving their cold storage around.
anyone who gets tricked by this sort of thing deserves to lose their money, quite frankly