And yet, they published that instead of figuring out what is possibly wrong with their gear which in turn may invalidate the rest of the their results. I'm not going to second guess their motivations though, they probably know what they are doing but it is interesting and deserves explanation.
Why publish this graph without running the measurements again though? Serious question since I’ve no idea about the effort needed to get this data.
I’ve got a small hope that they actually did and found the effect didn’t go away. They’ll still say ’equipment malfunction’, there isn’t any downside, only upside if it gets reproduced somewhere else.