Business

CAPEX, profit scenarios, roadmap, and gates.

Profit scenarios

Demand-led monthly P&L (a ~600-student community + walk-ins + events). CZK. A match is 20 min play + 10 min turnaround; a ticket is one player in one match, so a full match sells 12–16 tickets. All figures below are per ticket.

Three fixed scenarios. Marginal cost 15/12/10 CZK per ticket (pess./base/opt.); capex ~324k CZK (no DMX).

Driver	Pess.	Base	Opt.
Players / match	12	14	16
Tickets / mo	400	850	1,500
Price / ticket	200	250	320
Gross / mo (k)	80	212.5	480
Monthly opex (k)	95	90	85
Profit / mo (k)	−21	+112	+380
Payback (capex ~324k)	none	2.9 mo	0.9 mo

Capex ~324k (street prices Jul 2026, no DMX) = lighting ~99k + audio ~75k + tracking ~89k (wand fleet + anchors + network) + gateway PC / charger bank / hazer / props ~21k + obstacle build ~25k + install ~15k. That is the exact sum of the rows — there is no contingency inside it, and obstacle build and install labour are still estimates without a quote.

Profit as ticket volume grows, one line per price point. Hollow rings on the zero line = break-even (514 tickets at 200 CZK, 379 at 250 CZK, 275 at 320 CZK); filled dots = the demand each scenario assumes. Slope = contribution per ticket (price − marginal cost), so a higher price tilts the line up and pulls break-even left. The dashed line is the ~1,250-ticket capacity ceiling at 12 players and 12 h/wk.

Make-or-break: ~514 tickets/month covers 95k opex at the pessimistic price (200 CZK, 15 CZK marginal) — about 17 tickets/day, i.e. ~1.4 full 12-player matches. That demand threshold — not UWB capacity — decides viability. At 12 h/wk of opening hours the arena runs ~104 matches/mo, which is ~1,250 tickets at 12 players and ~1,660 at 16: pessimistic and base fit comfortably, while the optimistic 1,500 needs 16-player matches or longer hours.

Demand figures are internal scenarios, deliberately not anchored to public "FEC retention statistics" (aggregator numbers are unreliable — methodology-free content farms). The ~600-student community behaves like a club (higher visit frequency, lower price per ticket), not a family FEC trip. Validation path: primary research — operator calls and pilot sessions — not web statistics.

Roadmap & gates

Phase	Work	Duration	Exit gate
0	PoseProvider abstraction + doc truth	~1 wk	mostly done Gameplay decoupled from tracking source; a few doc-truth items open (RISK_AUDIT Phase A)
1	VLAN + boundary hardening + CI	~1–2 wk	partial Device secret, NVS creds, setup AP + shared web UI done; VLAN/AP + admin password at venue
2	D · Real UART parsing, first fix	2–4 wk	Gate D software ready — stable fix on hardware
3	A · Anchor-set assignment	~1–2 wk	hw pending Code done; validate moving wand
4	B · TDMA (WiFi-sync → beacon)	2–4 wk	Gate B Code done — 16 wands @ ≥8 Hz on hardware
5	Capacity measurement + tune	~1 wk	Documented ceiling ≥ 16
6	Arena build-out (lighting + audio)	4–6 wk	After Gate 5; <100 ms end-to-end; v1 cameraless
6b	Bench G0–G3 + v2 sparse retro optical	parallel	after Gate D 4–6 yaw-well cameras if gates pass
7	Ops / compliance / pilots	parallel	Ready for first paid session

v1 launches cameraless (UWB + IMU + re-zeros). Retro band optical v2 after Gate D + G0–G3. Spec: docs/superpowers/specs/2026-07-21-retroreflective-optical-tracking-design.md. Laser-tag vest hits out of scope.