Sixty ways to find nothing

Section 1

What a miss looks like from the inside

A miss arrives in one of three shapes, and they do not overlap. The brief either wrote the gap down, or stopped and put a question back to the owner, or recorded nothing at all. Crossed with whether anything was examined first, that is the whole anatomy.

• The brief named the gap in writing and said what is missing.50
  • 40examined candidates
  • 10examined nothing

  Risks in this group: tariffs, rules and policy costs (14), currency on imports and exports (9), local, event-driven demand (8)

• The brief stopped and put a question back to the owner. Candidates had been examined; the risk was not stated precisely enough to choose between them.9
  • 9examined candidates
  • 0examined nothing

  Risks in this group: weather that stops the day (9), food and ingredient costs (1)

• The brief recorded neither a gap nor a question. The miss left no note.1
  • 0examined candidates
  • 1examined nothing

  Risks in this group: promotions that pay out (1)

Read the second column first. In 11 of the 60 misses the exchange offered nothing close enough to open. In the remaining 49, there were candidates on the table, 203 of them, and every one was turned down.

Section 2

Why the candidates were turned down

Across the corpus, 250 contracts were examined and 219 were rejected with a written reason. The reasons are near-unique text: 207 of them are different strings. They are grouped below by an ordered set of patterns, first match wins. The full rule list is printed at the end so the grouping can be argued with.

• The market measures a different thing entirely98
  • 44.7%of rejections
  • 42briefs affected
  • Central Pacific storm naming has nothing to do with New Orleans weather risk.
  • Fertilizer is a farm input cost, not the food price you actually pay at the bar.
• The market watches a different stretch of time39
  • 17.8%of rejections
  • 30briefs affected
  • Counts total Eastern Pacific storms for the year, not a single local landfall this quarter.
  • Same wrong peril and wrong season as the other tornado count line.
• The market measures somewhere else36
  • 16.4%of rejections
  • 16briefs affected
  • Counts Central Pacific storms overall, wrong basin and wrong trigger for a New Orleans disruption.
  • Counts storms in the Eastern Pacific basin, not rain or storms hitting Tampa.
• The line is priced so close to certain that cover costs about what it pays16
  • 7.3%of rejections
  • 16briefs affected
  • Priced almost certain to trigger, so it offers little real protection for the cost.
  • Prices near-certain at this level, so it offers almost no real protection for the cost.
• The line is priced so far out that it would almost never pay12
  • 5.5%of rejections
  • 12briefs affected
  • Priced too near-impossible to pay out much if rates only stay moderately high.
  • This price line is priced near-impossible, so it offers almost no real protection.
• Right subject, wrong trigger level11
  • 5.0%of rejections
  • 8briefs affected
  • Sets the rain bar too high, so it pays only in rare extreme months and misses smaller disruptive storms.
  • This line is far above the stated $140 spike, so it would rarely pay when the owner's cost risk hits.
• A different line on the same ladder was taken instead7
  • 3.2%of rejections
  • 6briefs affected
  • Lower price line pays too easily, offering weaker protection than the $1100 line for the same cost spike.
  • Not selected in this saved analysis.

219 of 219 notes fall under a rule and 0 do not. That coverage is descriptive rather than predictive: the rules were written by reading these same notes, which the methods block states plainly.

Section 3

Which risks found nothing

Sorted by how often the risk was detected. Of the 9 risk types the corpus names, 3 matched a contract zero times. The figure on the right of each row is the briefs that found a contract over the briefs where that risk was detected.

• Weather that stops the day0/15
  • 15times detected
  • 15briefs left uncovered
  • 4gaps written down
• Tariffs, rules and policy costs1/15
  • 15times detected
  • 14briefs left uncovered
  • 14gaps written down
• Borrowing costs4/10
  • 10times detected
  • 6briefs left uncovered
  • 6gaps written down
• Currency on imports and exports0/9
  • 9times detected
  • 9briefs left uncovered
  • 9gaps written down
• Local, event-driven demand0/8
  • 8times detected
  • 8briefs left uncovered
  • 6gaps written down
• Food and ingredient costs3/8
  • 8times detected
  • 5briefs left uncovered
  • 4gaps written down
• Metals and materials costs1/7
  • 7times detected
  • 6briefs left uncovered
  • 6gaps written down
• Fuel and energy costs1/6
  • 6times detected
  • 5briefs left uncovered
  • 4gaps written down
• Promotions that pay out1/5
  • 5times detected
  • 4briefs left uncovered
  • 3gaps written down

Section 4

The gaps, in the product’s own words

When a brief could not cover a risk it wrote the gap down under a name. These are those names, and how many times each was written.

• tariffs and policy: no clean market14
• currency swings: no clean market9
• financing costs: no clean market6
• local, event-driven demand: no clean market6
• materials costs: no clean market6
• food and ingredient costs: no clean market4
• fuel and energy costs: no clean market4
• weather: no clean market4
• promotional payouts: no clean market3

9 further briefs never got as far as a gap. They stopped and asked the owner a question instead. Every one of them asked the same question: What weather would hurt the business most: heavy rain, extreme heat, cold, or another condition? Those misses are not the exchange’s doing. They are what happens when a risk is stated too loosely for any contract to settle on it.

Section 5

When it did fit, how well

The 11 briefs that selected a contract each recorded how loosely the contract is tied to the business result. Every one of them recorded a mismatch.

• The market and your business result can move differently.10
• The contract is decided by the public event, not by your business's actual loss.1