Little Red-Eyed Riding Actuary Assumptions underlying this fairy tale model Wolves can talk and are pretty good at mimicry. Actuaries engage in conversation with strangers, even if those strangers are talking wolves. Wolves are frightened of actuarial recruiters. Wolves are able to swallow people whole. Actuaries do not induce vomiting when eaten. A wolf’s stomach can stretch enough to fit two people (“best fit”). The aforementioned swallowed people don’t die. Actuarial recruiters are skilled in the use of axes. Once upon a timeline, in a life insurer’s office, there worked an actuarial intern. She was called “Little Red-Eyed Riding Actuary” because she was new and worked so many hours to please her team that she had bloodshot eyes. One day her manager, having heard her grandma was ill, said to her, “Go see how your grandmother is doing, for I hear that her force of mortality is increasing exponentially. This may be the ideal opportunity to see if she’s interested in buying some long-term care or critical illness cover funded by an equity release.” Little Red-Eyed Riding Actuary set out immediately to go to her grandmother, who lived in another village. As she was walking through the woods, she met a wolf who wanted to eat her up, but he dared not because some actuarial recruiters had already latched onto the scent of Little Red Eyed Riding Actuary and were closing in. The wolf asked her where she was going. The young actuary did not realise that it was dangerous to engage in small talk with anyone, let alone wolves. For, if she’d carried out some basic exploratory data analysis of fairy tales, she would have quickly discovered that wolves are statistically very likely to be the villains of the story. “I am going to see if grandmother is interested in purchasing long-term care or critical life cover.” “Does she live far off?” asked the wolf. “It is the first house on the right beyond the mill,” answered Little Red-Eyed Riding Actuary, breaching GDPR. Now Little Red-Eyed Riding Actuary had calculated the quickest route to get there, considering the effect of the rising terrain on her rate of walking. However, the wolf pointed out to her that loggers had been working on a different path and she would maximise her log-likelihood of getting there if she were to take it. This seemed sensible to the girl, and she took that path while the wolf ran as fast as he could along the shortest route to the grandmother’s house and knocked on the door. “Who’s there?” “Your grandchild,” replied the wolf, trying hard to make his voice sound boring enough to be mistaken for an actuary. “Lift the latch,” called out the grandmother, “I am in a morbid state, and cannot get up.” The wolf lifted the latch; sprang in and ate her up in one bite. He then shut the door, dressed himself in her nightcap, got into the grandmother’s bed, and used a constant arrival rate assumption to model the waiting time before Little Red-Eyed Riding Actuary knocked at the door. “Who’s there?” Little Red-Eyed Riding Actuary was at first concerned upon hearing the deep voice of the wolf, but then realised that 95% of ill people’s hoarse voice confidence intervals would still include this sound. She answered, “It is your grandchild, Little Red-Eyed Riding Actuary.” The wolf softened his voice as much as he could and said, “Lift the latch and come in.” When Little Red-Eyed Riding Actuary came in, she saw her grandmother lying in bed with her cap pulled far over her face, and looking like a definite outlier to the Grandmother’s appearance distribution. Nevertheless, she started with a null hypothesis that it was indeed her grandmother and carried out a test based on a sample of her grandmother’s appearance. “Oh grandmother,” she said, “What big ears you have!” “All the better to hear you say that I’m still within three standard deviations of the mean, my child.” “But, grandmother, what significantly big eyes you have!” “All the better to see that a sample size of only one grandma means your standard error is too large to draw any conclusion, my child.” “But, grandmother, what a big chamber pot you have!” “All the better for my large p-value, my child.” “Oh! But, grandmother, what big teeth you have!” “All the better to shred any ‘correlation is causation’ arguments you may have.” And so Little Red-Eyed Riding Actuary became convinced that she had insufficient evidence to reject her null hypothesis at the 5% level. But then she noticed the fat tail distribution of her grandmother and gasped, “Oh! But, grandmother, what a significantly big tail weight you have!” And realising that his game was up, the wolf swallowed up Little Red-Eyed Riding Actuary in one gulp! “I appear to have made a Type 2 error,” sighed Little Red-Eyed Riding Actuary from within the wolf’s stomach and vowed never to talk to strangers again. Now the wolf was feeling quite nauseous after eating the bland actuary, for she was not as good to eat as his tastiness model had predicted. So he lay down again in the bed whilst he updated his model, but instead fell asleep, and began to snore loudly. An actuarial recruiter who was “just passing” the house looking for someone, literally anyone, to fill a post, heard the snoring and went in. The recruiter discounted the strange appearance of the grandmother at a rate of 5% pa because he got his commission, regardless. However, when the recruiter noticed the posterior distribution of the wolf’s belly, he used an axe to split the wolf into its principal components. Little Red-Eyed Riding Actuary was so grateful to be alive and desirous to do a more boringly safe job, she readily accepted the position offered by the recruiter. The grandmother now realised the importance of life cover and purchased every product available from the actuary. And the Profession was delighted as they could now collect another year’s worth of membership fees from Little Red-Eyed Riding Actuary.
This cracked me up. The wolf also made a type 2 error, assuming his null hypothesis was actuaries are good to eat.