So in general, you'd expect an inverse correlation with BTC price?
It was supposed to be a linear relation if you take a look at the source. It seems like the graph was made by another user and they deliberately showed an inverse relation graph. The "drop" in the MDIA on the graph above must be interpreted as a higher point of MDIA and vice versa.
I don't know. The chart at the link supplied does have that sharp drop at the end of 2017, unless I'm reading it incorrectly - the lighter of the two blue colours is the MDIA?
During accumulation periods, like 2017, the MDIA reaches very high levels. On the other hand, sharp drops in MDIA point to BTC purchased at "expensive" rates being sent back to exchanges, and such events seem to be followed by price drops.
But how can MDIA
increase during high accumulation? If we mean new money coming in from fiat, or even just into new bitcoin addresses, then that's a
reduction in MDIA, surely? More significant if the price is rising, but still true regardless of price. Even quoting from the site: "Every price top so far was accompanied by a drop in the mean coin age and with a significant drop of mean dollar age. During the top on January, 2018 the dollar age went down to just 9 weeks."
A sharp drop in MDIA has to be tied to new money coming in, which is generally when prices are rising, so an inverse relationship between MDIA and price. And then with a long bear market and no new money coming in, the MDIA rises at almost a day per day. That's all it's saying, isn't it? I can see that the metric is of some interest, but I can't see any profound insights.