It’s also not like there’s any requirement for reporting energy figures by miners. 2.8x seems, to me, a very good estimate for something with large errors inherently.