samber/dev-event-organizer-skills / skills/event-registration-forecast
event-registration-forecast
VerifiedForecast final registrations for a technical event from its sign-up pace curve, with a bundled Python 3 script, while registration is still open - prior-edition or generic reference curves aligned on days-out, a low/mid/high band, milestone tripwires declared in advance, the commit figure at each contract lock date, when the forecast crosses capacity, and attendance by ticket type. Use whenever asked how many people will end up registering or attending, to project this year from last year's curve, what headcount to commit before sign-ups close, or when registration will hit its cap. Do NOT use for the demand verdict - use samber/dev-event-organizer-skills@event-market-fit - or overbooking ratios and food counts once sign-ups close - use samber/dev-event-organizer-skills@event-no-show-management.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 20.2 KB |
| assets/curve-free-hackathon-target.csv | other | 63 B |
| assets/curve-pooled-lead-time.csv | other | 80 B |
| assets/ticket-types-example.csv | other | 183 B |
| evals/evals.json | config | 49.0 KB |
| references/attendance-by-ticket-type.md | markdown | 2.6 KB |
| references/lock-dates-and-commit-rules.md | markdown | 5.0 KB |
| references/reference-curves.md | markdown | 4.6 KB |
| references/worked-forecast.md | markdown | 4.4 KB |
| scripts/forecast-registrations.py | script | 17.4 KB |
references/worked-forecast.md
View on GitHub# Worked forecast
A paid two-day conference, second edition, room for 500. Last edition finished at 500 registrations net of comps. The team plans 550 this year, and break-even needs 350.
## Inputs
`last-edition.csv` (days_out, cumulative): 120,20 / 90,45 / 60,110 / 45,170 / 30,260 / 14,380 / 7,430 / 0,500.
`actuals.csv`: 120,25 / 90,55 / 60,120 / 45,165. Speaker announcement at 60 days out both years; early-bird close at 50 days this year, 45 last year.
`ticket-types.csv` (type, basis, amount, show_rate, label): paid, share, 0.85, 0.90 (sourced paid band, low end); student, share, 0.15, 0.75 (self-set); speakers, count, 40, 0.97 (self-set); sponsor staff, count, 30, 0.70 (self-set).
Caterer: guarantee 7 days out, can add 5% up to two days out. Hotel block: rate code expires 30 days out, 80% usage guarantee.
## Command
The early-bird close moved five days earlier, so last edition runs twice: as is, and shifted `@+5` so its early-bird spike sits on this year's date.
```
python3 scripts/forecast-registrations.py --actuals actuals.csv \
--reference last-edition.csv@+5 --reference last-edition.csv \
--reference assets/curve-pooled-lead-time.csv \
--target 550 --capacity 500 --threshold 350 \
--checkpoint catering:7 --checkpoint hotel-block:30 --error-band 0.1 --types ticket-types.csv
```
## Output, read line by line
```
Projecting from: 165 registrations at 45 days out
- last-edition.csv shifted +5d (now at 0.82x its count at this days-out): expected share 40.0%,
pace index 75, multiplicative 412 -> uses 412
- last-edition.csv (now at 0.97x): expected share 34.0%, pace index 88, multiplicative 485 -> uses 485
- curve-pooled-lead-time.csv: expected share 42.8%, pace index 70, multiplicative 386 -> uses 386
Projected final registrations: low 347 / mid 412 / high 534 (spread, widened by +/-10%)
Checkpoints: catering (7 days out): 297-459 on hand
hotel-block (30 days out): 188-278 on hand
Threshold 350: mid clears, low does not
Total expected attendance: 364 / 422 / 528
Capacity 500 (attendance incl. comps): low 73%, mid 84%, high 106%
```
- **The shift matters.** Unshifted, last edition says 485; aligned on the early-bird date it says 412. The 45-day reading already holds this year's early-bird spike, which last edition only reached at 45 days, so the unshifted ratio flatters this year. Lead with the shifted run, keep the unshifted one as a reference, and let the pooled curve set the low end. The mid lands on the shifted run because last edition counts twice in the median; say so.
- **Pace index 70-88 against the 550 plan.** The event is behind a plan that assumed growth and roughly level with, or below, last year. That is a gap against the target; hand it to `event-market-fit` as a signal only if the 30-day reading lands under 188.
- **Declare the 30-day range now: 188-278 on hand.** The reading after an early-bird close carries the lull; a figure inside the range is the deadline working, not demand falling.
- **Threshold: mid clears, low does not.** No escalation yet. Re-run at the go/no-go review before the 30-day cancellation-fee step; if the mid then misses 350, the escalation is promoted.
- **Capacity.** High-end attendance, 528 with 60 comps attending, passes the 500-person room by 6%. A paid event takes no overbooking, so paid registrations close at 440, the 500-person room minus the 60 comps expected in it: tell `event-no-show-management` and `event-venue-sourcing` now, and flag the larger-room option if the 30-day reading lands above 278.
## Lock-date table produced
| Commitment | Date | Commit | Band end | Reason |
| --- | --- | --- | --- | --- |
| Hotel block | 30 days out | 291 room-nights (low end, 40% out of town, 2 nights) | low | 80% usage guarantee; hotel confirmed it can add rooms |
| Catering, day one | 7 days out (provisional now, re-run on the day) | 503 covers, +5% flex to 528 | high | A short lunch costs more than a spare cover; 528 / 1.05 keeps the flex for a late surge |
| Catering, day two | 7 days out | Unknown: no per-day check-ins from last edition | - | Ask the caterer for a later day-two guarantee and count day-one check-ins first; never copy day one |
## Negative example: the straight line
```
// ✗ Bad - 165 sold in 75 days of sale, 45 days left: 165 / 75 x 45 = 99 more, final ~264
```
The straight line assumes registrations arrive evenly. Last edition took 330 of its 500 in the final 45 days. It would have read this event as missing break-even by 86 and triggered a downgrade the curve never justified.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |