I might be misunderstanding you, but given my interpretation of this post it seems like an unfair comparison?
“Chartist haters” don’t usually claim at time that the trend will break at time , but rather that it will break at some later point, pushing back against “chartist lovers” who might seem to reason in a way like “If it’s still going up from to , then it will probably keep going from to . And if it’s still going up from to , then it will probably keep going up from to . And so on all the the way to [or something].”, rather than lawfully multiplying conditional probabilities.
Well, it actually depends on how you partition the time steps. Is the next day? Year? Decade? Millenium? But you talk about “short-term chartist haters”, and I think that the “hating” happens over longer horizons than the short-term of those who “love” “short-term charting”.
Note that I specified the “short-term” chartist position carefully specifically to delineate a group away from objections like yours.
I don’t think many chartist haters are typically as smart or humble as you imply they are. During COVID there were waves upon waves of people who kept insisting that the virus will break any day now against the clear exponential and calling us naive, and they seemed to not update much on being wrong just days before.
Well, it actually depends on how you partition the time steps. Is the next day? Year? Decade? Millenium?
The way I’ll think about it is: if a trend has been true for N years, where N is between 10 and 100, N+1 probably refers to years rather than days. If it’s been true for 0.5 years, N+1 probably doesn’t refer to years.
One reason your choice of unit matters less than you might think is that Laplace’s law of succession is pretty helpful here, and the numbers converge nicely (1 year after 100 years gives you similar number to 365 days after 100*365).
I might be misunderstanding you, but given my interpretation of this post it seems like an unfair comparison?
“Chartist haters” don’t usually claim at time
Well, it actually depends on how you partition the time steps. Is
Note that I specified the “short-term” chartist position carefully specifically to delineate a group away from objections like yours.
I don’t think many chartist haters are typically as smart or humble as you imply they are. During COVID there were waves upon waves of people who kept insisting that the virus will break any day now against the clear exponential and calling us naive, and they seemed to not update much on being wrong just days before.
The way I’ll think about it is: if a trend has been true for N years, where N is between 10 and 100, N+1 probably refers to years rather than days. If it’s been true for 0.5 years, N+1 probably doesn’t refer to years.
One reason your choice of unit matters less than you might think is that Laplace’s law of succession is pretty helpful here, and the numbers converge nicely (1 year after 100 years gives you similar number to 365 days after 100*365).