3 ms·
I think Kahneman has a name for this common fallacy in this book -- it's the one wherein small samples always appear to change radically because they are... uhm
by eliben 13y ago
I think Kahneman has a name for this common fallacy in this book -- it's the one wherein small samples always appear to change radically because they are... uhm... small.
- jacques_chester 13y agoIt's an artifact of small figures and small sample sizes generally. Anything which grows geometrically will grow progressively more slowly over time when expressed as a percentage.
- adventured 13y agoIncorrect. If you were right, then every provider that was small would be skyrocketing up the usage charts. However, that's numerically impossible. Many smaller providers simply die off or never gain any traction. This time next year, they'll likely be in the top 15 to top 25 on that same list.
- eliben 13y agoPrecisely. When a smaller provider goes from 1000 to 50 users, it's a "ZOMG 95% drop!" while only 950 users left, line noise for bigger providers. You're adding another example to my point. Naturally I didn't say that all small numbers always change drastically. But those that do change, appear to be more drastic than larger numbers.