I don’t like the use of the terms “white swan” and “black swan” here.
The original use was derived from the way that people who hadn’t been to Australia would (reasonably) have thought that all swans were white, and would have found the existence of black swans a completely unpredicted surprise, outside the scope of their mental models of the world. A “black swan” in this sense is just something very surprising, unprecedented, beyond the normal variation one’s taking into account. (And a “white swan” would be the usual case: business as usual, nothing interesting happening.)
But in this article and its predecessor, it seems that “black swan” is being used to mean “unpleasant surprise” and “white swan” to mean “pleasant surprise” or “opportunity” or something.
This seems to me a much less useful bit of terminology than “black swan” originally was, and it would make me sad if it were to catch on.
Isn’t it taken directly from Nassim Taleb’s The Black Swan, where he advocates defenses against ‘black swans’ but conversely deliberately opening yourself up to ‘white swans’ through tactics like moving to cities?
It’s some time since I read that book. I don’t remember him using “white swans” to mean good out-of-model events, but that might just be memory failure on my part. I remark that here you can see Andrew Gelman (who is very clever) having just read the book and clearly using “black swan” to cover good things and “white swan” to mean “smaller within-model predictable things”.
Alright, so I redownloaded Black Swan off Libgen and have been browsing through it. Taleb is a little confusing, but looking at passages, I think roughly Taleb defines things as:
“Black swan”: extreme unpredicted events; this unpredictability can be due to models or calculations that fail to incorporate power laws of appropriate exponents, or it can be due to model uncertainty / Knightian uncertainty / closed-universe assumption
“Gray swan”: extreme events which can be modeled by an appropriate fractal/power-law model, although they only can give very vague predictions.
“White swan”: common (not rare) events predicted by one’s model, excluding the unpredicted black-swans and the possibly-predicted-but-still-rare gray swans
This overlaps with his ‘Mediocristan’/‘Extremistan’ dichotomy; he specifically rejects there being any moral or desirability connotation to black vs white swan, other than commenting on the unfairness and randomness of black swans (whether they’re either positive or negative for the affected individuals, be they J.K. Rowling or someone boarding a flight on 9/11).
I thought he had a table of recommendations including things like living in cities, but I can’t seem to find it skimming Black Swan, so perhaps I saw that in some of his writings since then.
So, I guess these articles are misusing ‘white swan’ after all.
That fits well with my admittedly hazy recollection, and suggests that the usage here (“swan” = “unexpected event”, “black” = “bad”, “white” = “good”) is quite different from Taleb’s (“swan” = “event”, “black” = “out-of-model”, “white” = “normal”).
I thought he had a table of recommendations including things like living in cities, but I can’t seem to find it skimming Black Swan, so perhaps I saw that in some of his writings since then.
Probably—at some point Taleb was writing about optionality and how having some is a very good thing (you expose yourself to the volatility but cut off the left tail). Like you I don’t remember the sources, but I think it was a few years after the Black Swan.
Well, he seems to denote predictability with the colours, which was the point. Making white swans indicate rare good events makes no sense. The Wikipedia page suggests he uses the term black swan also for good events.
“Isn’t it taken directly from Nassim Taleb’s The Black Swan”
Right.
“Making white swans indicate rare good events makes no sense.”
Actually, you could be right, but that’s how I’m using it. I don’t have my copies of Taleb’s books in front of me, but I’m pretty sure he uses the terms the way I’m using them.
I don’t like the use of the terms “white swan” and “black swan” here.
The original use was derived from the way that people who hadn’t been to Australia would (reasonably) have thought that all swans were white, and would have found the existence of black swans a completely unpredicted surprise, outside the scope of their mental models of the world. A “black swan” in this sense is just something very surprising, unprecedented, beyond the normal variation one’s taking into account. (And a “white swan” would be the usual case: business as usual, nothing interesting happening.)
But in this article and its predecessor, it seems that “black swan” is being used to mean “unpleasant surprise” and “white swan” to mean “pleasant surprise” or “opportunity” or something.
This seems to me a much less useful bit of terminology than “black swan” originally was, and it would make me sad if it were to catch on.
Isn’t it taken directly from Nassim Taleb’s The Black Swan, where he advocates defenses against ‘black swans’ but conversely deliberately opening yourself up to ‘white swans’ through tactics like moving to cities?
It’s some time since I read that book. I don’t remember him using “white swans” to mean good out-of-model events, but that might just be memory failure on my part. I remark that here you can see Andrew Gelman (who is very clever) having just read the book and clearly using “black swan” to cover good things and “white swan” to mean “smaller within-model predictable things”.
Alright, so I redownloaded Black Swan off Libgen and have been browsing through it. Taleb is a little confusing, but looking at passages, I think roughly Taleb defines things as:
“Black swan”: extreme unpredicted events; this unpredictability can be due to models or calculations that fail to incorporate power laws of appropriate exponents, or it can be due to model uncertainty / Knightian uncertainty / closed-universe assumption
“Gray swan”: extreme events which can be modeled by an appropriate fractal/power-law model, although they only can give very vague predictions.
“White swan”: common (not rare) events predicted by one’s model, excluding the unpredicted black-swans and the possibly-predicted-but-still-rare gray swans
This overlaps with his ‘Mediocristan’/‘Extremistan’ dichotomy; he specifically rejects there being any moral or desirability connotation to black vs white swan, other than commenting on the unfairness and randomness of black swans (whether they’re either positive or negative for the affected individuals, be they J.K. Rowling or someone boarding a flight on 9/11).
I thought he had a table of recommendations including things like living in cities, but I can’t seem to find it skimming Black Swan, so perhaps I saw that in some of his writings since then.
So, I guess these articles are misusing ‘white swan’ after all.
That fits well with my admittedly hazy recollection, and suggests that the usage here (“swan” = “unexpected event”, “black” = “bad”, “white” = “good”) is quite different from Taleb’s (“swan” = “event”, “black” = “out-of-model”, “white” = “normal”).
Probably—at some point Taleb was writing about optionality and how having some is a very good thing (you expose yourself to the volatility but cut off the left tail). Like you I don’t remember the sources, but I think it was a few years after the Black Swan.
Taleb doesn’t seem to be terribly consistent about non-black swans...
Well, he seems to denote predictability with the colours, which was the point. Making white swans indicate rare good events makes no sense. The Wikipedia page suggests he uses the term black swan also for good events.
“Isn’t it taken directly from Nassim Taleb’s The Black Swan”
Right.
“Making white swans indicate rare good events makes no sense.”
Actually, you could be right, but that’s how I’m using it. I don’t have my copies of Taleb’s books in front of me, but I’m pretty sure he uses the terms the way I’m using them.
Did you check out the article Lumifer linked to? Perhaps Nassim changed his mind.
Especially given that there already is a term specifically for positive black swans, namely “windfall”.
But there definitely should be a term specifically for negative black swans too.