This lecture contains a similar graph on asset prices - he uses returns from the S&P for the period 1871-2013. The book really is worth a read too, it's more accessible than the lecture or slides might indicate.

https://www.gurufocus.com/news/336523/speculative-asset-pric...

https://www.nobelprize.org/uploads/2018/06/shiller-lecture-s...

The comparison is between the actual stock prices against the constant discount rate of subsequent real dividends.

Day to day prices are unforecastable, and it's not really understood what drives pricing in the market.

> The comparison is between the actual stock prices against the constant discount rate of subsequent real dividends.

The EMH states that the market incorporates all currently available information, but future dividends are known only in hindsight. Prices that fluctuate around future earnings therefore are not in conflict with efficient markets at all.

I'd recommend the book if you haven't read it, I don't think my poor summaries do it justice, and it does address some of your objections from memory, and also doesn't attempt to throw out EMH, just questions whether it fully explains market behaviour.

Yes the chart compares measured return (divs) against projected return (prices) so it is not a prediction machine but I think it does illustrate well that market prices are usually far from true value.