Travise
PricingSign inStart free
Travise

A second read for active traders.

Account
  • Pricing
  • Sign in
  • Create account
Learn
  • Research
  • Compare
  • Trading questions
  • Free tools
Help & legal
  • Support
  • FAQ
  • Privacy
  • Terms
  • Cookies
  • Sub-processors

Travise reviews trades you have placed and checks a planned trade against rules you set yourself. It does not tell you what to trade. It is not investment advice, not a signal service, and not a recommendation to buy or sell. You bear all trading risk.

Travise live
© 2026·e4a7434f
← Research

Methodology

The margin of safety was a ratio, not a mood

Graham computed a number before he committed. The phrase has since drifted into meaning something like caution. Here is what he actually calculated, and what a trader can honestly borrow from it.

Maxwell Norman · September 2, 2026 · 5 min read

"Margin of safety" has drifted into meaning something like caution, or leaving yourself room. That is not what Benjamin Graham meant. He meant a number you compute before you commit, and the trading version is the same arithmetic pointed at a losing run.

The phrase gets used now the way "risk management" gets used: as a mood rather than a measurement. Somebody says they trade with a margin of safety and what they mean is that they feel careful. Graham would not have recognised that. In his hands it was a ratio, worked out on paper, before the money moved.

What he actually computed

His clearest example is a bond, and it is worth walking because the shape is what transfers. A company earns, say, five times its fixed charges. The margin of safety is that multiple. Earnings can fall by eighty percent and the interest still gets paid.

Notice what that is and is not. It is not a forecast of what earnings will do. It is a measurement of how wrong the forecast can be before it costs you. You compute it from figures already on the page, and you compute it before buying, not after the price moves against you.

Not a forecast of what happens. A measurement of how wrong you can be before it costs you.

Graham said the point out loud, and it is the sentence worth keeping: the function of the margin of safety is, in essence, that of rendering unnecessary an accurate estimate of the future. The buffer exists so you do not have to be right about what happens next.

The trading version

A trader is not buying a bond with a fifty-year earnings record. There is no intrinsic value to discount to, and no coverage ratio to compute. The idea has to be re-pointed at something a trader actually has, and what they have is a win rate and a size.

Run it the same way. Take the losing streak that is normal at your win rate. Work out what it costs at the size you trade. If the answer is a number you can absorb and keep trading afterwards, you have a margin of safety. If it is a number that ends you, you do not, however careful you feel.

The property that carries across is the important one: you do not need to predict the streak. You need to be sized so that the streak arriving is an event rather than a conclusion.

✦Travise Read (AI-generated)

What Travise computes here

Two figures, both from your own closed trades. The losing run that is normal at your measured win rate, and the longest run you have already lived through. When the second is longer than the first, the second one leads, because it is a fact and the other is a model. The arithmetic also uses the pessimistic end of your win rate rather than the middle, which is Graham's own principle applied to its own input.

What does not survive the transfer

Most of it, and pretending otherwise would be the same kind of borrowed-authority move the phrase already suffers from.

Graham's margin of safety is always dependent on the price paid. Buy the same business cheaper and the margin grows. There is no equivalent for a trader holding for three hours: your entry price sets your risk, but it does not buy you a discount to some underlying worth, because you are not holding long enough for worth to assert itself.

Nor does the diversification half carry. Graham's buffer works partly because it is applied across a basket, where individual mistakes cancel and the aggregate wins. Twenty correlated intraday positions in the same index are one position wearing twenty tickets.

So this is a borrowing, not a lineage. What comes across is the shape: compute the buffer before you commit, from numbers you already have, and size so that being wrong repeatedly is survivable. What stays behind is nearly everything else Graham wrote, and any page telling you otherwise is selling you a dead man's endorsement.

Travise is the senior trader’s read on every trade you take. $0 today. Cancel before day 7.

Grade my first 7 trades free