Evidence and reading

Four resources

What this book wants to help you keep more of is not only lifespan; there are four things in all:

• Lifespan: live longer, and die less often of things that could have been avoided
• Time and energy: the time you are alive isn't spent on things with no return, and less of each day's attention and physical energy is drained for nothing
• Money: waste less money, and spend it where the return is certain
• Personal freedom: don't land yourself in a detention center (拘留所) or a pretrial detention center (看守所) just because you didn't know about a red line

Every piece of advice answers two questions: what you spend (money / time / energy / willpower), and what you get back (change in all-cause mortality / a drop in a specific cause of death / time and energy saved / money saved / security and personal freedom). Items are ranked by value for cost, not by category: those that cost almost nothing and bring back a large benefit come first.

Who receives the benefit of a piece of advice comes in tiers. From highest to lowest by “how likely this benefit is to come back to you yourself in the future”: ① you yourself; ② your spouse and immediate family (parents, children, paternal and maternal grandparents, paternal and maternal grandchildren); ③ friends, colleagues and other relatives: a reciprocal relationship, where help you give may come back in the future; ④ strangers: the lowest tier, but not zero: the chance of a return is small, and since you don't know the other person's character, there is also the side of being extorted, having the blame turned on you, or being retaliated against. Different tiers are not added together, and where tier ④ comes up, the benefits and the risks are written together.

Read the first-aid section this way: of 38,227 out-of-hospital cardiac arrests in China, 79.2% happened at home, so learning chest compressions is first of all for doing them on your own family; rules such as “if you see someone drowning, don't go into the water yourself” and “if you come across a fight, don't step in to pull people apart” are self-protection rules in themselves, guarding against you going from bystander to second casualty. Whether to step in for a stranger is your own trade-off; the items spell out the liability-exemption provisions, the self-protection steps and the risks, and do not count it for you as a reason you must act.

Mortality figures, time-and-energy figures, money figures and legal consequences are each counted separately and are not converted into one another. These four correspond to the four “what you get back” options on the upstream search page, and are not ranked against one another.

Evidence grades

Every piece of advice is marked with an evidence grade:

Grade	Meaning
A	There is a specific number to check, and the source is a meta-analysis pooling several studies, a large cohort that followed many people for many years, or a randomized trial (RCT); it can say how much something fell (HR, RR, percentage drop)
B	Supported by research, but no exact number can be given; or only a small sample or a single study holds it up
C	The author's own experience, or commonly accepted practice, with no direct research literature

Of the book's 676 items, 439 are grade A, 182 grade B and 55 grade C; in addition, 72 items are marked contested, and 3 places are marked TODO, to be verified. Contested grade A/B items are marked “contested” and list the opposing evidence. All sources cite only original literature (journal papers with a DOI or PubMed link, or reports from official bodies such as WHO/CDC/the National Bureau of Statistics), never secondhand retellings. For cases the authorities do not make public, a few items cite original news reports bylined by a journalist, used only to show that “this kind of thing has happened”; the Sources field attaches a web archive, and the Notes say “media reports only”. Uncertain numbers are marked “to be verified”.

Grade A only means “there is a specific number and the source can be checked”; it does not mean the number is necessarily causal. Grade A includes both randomized trials and studies that only followed and recorded people without randomizing them; the latter cannot tell whether the thing itself had an effect or whether the people who do it were healthier to begin with. To see which kind, look at the Benefit field: one that says “randomly split into two groups” is the former, and one that says “only observed, not randomized” is the latter. For legal and policy items, grade A means the original text of a statute or official document is cited.

The percentages in the Benefit field and in In plain terms mostly mean so many tenths lower than for people who didn't do it, not so many fewer out of every hundred people. With the same 12% reduction, older people, whose risk is high to begin with, avoid far more deaths than young people, whose risk is very low to begin with. So first look at which people In plain terms is talking about, then look at the percentage.

Value-for-cost tiers

The evidence grade answers only “can this number be trusted”, not “is this worth doing”. So each item is also marked with two more things: benefit size (how big the benefit is) and outcome type (what kind of thing you get back). The upstream search page combines these two with the three costs into a value-for-cost tier:

Dimension	Values	How it is set
Outcome type	Lifespan / Money / Time and energy / Personal freedom	By what the item mainly gets you back. Different outcome types are not compared: “all-cause mortality down 12%” and “500 yuan saved a year” are not on the same ruler and can't be ranked against each other
Benefit size	Large / Medium / Small	As far as possible, follows the item's own “Benefit” field and applies cutoffs set in advance, not gut feeling: for lifespan, by how many percent it fell (≥20% is large, 10–20% medium, and <10%, or measured only on an intermediate marker and not on the final outcome, small); for money, by amount (on the order of ten thousand yuan is large, hundreds to thousands medium, tens of yuan small); for personal freedom, by consequence (avoiding criminal liability is large, avoiding detention or an administrative penalty medium, avoiding a civil dispute small); for time and energy, by how much is saved (hours saved a day is large, hours a week medium, a one-time saving small)
Value for cost	Very high / High / Standard	Large benefit with all three costs at zero = very high; large benefit with costs that aren't high, or medium benefit with zero cost = high; everything else = standard

Of the book's 676 items, 114 (17%) are very high value for cost, 303 (45%) high and 259 (38%) standard. The large number in the middle tier is intentional: the benefit size underneath has only three levels to begin with, large, medium and small, and cutting it any finer would be pretend precision.

This tier is the author's own judgment, not evidence; by this book's standard it counts only as grade C itself; it and the evidence grade are two separate things, and neither affects the other. An item can have grade A evidence but only standard value for cost (the shingles vaccine has a phase III RCT showing 97.2% efficacy, but two doses cost three to four thousand yuan, and shingles is rarely fatal), or only grade C evidence but very high value for cost (send your itinerary to your family before leaving the country). “Standard” does not mean you shouldn't do it: every item in the book is recommended; it's just that at this tier you have to weigh for yourself whether the expense is worth it.

Value tiers within each outcome type

Counts of items per value tier, shown separately for each outcome type because the guide does not compare tiers across outcome types.