dcf-quick

English · 简体中文 · Site

A one-page DCF that tells you how much to buy — not just what it's worth.

Most DCFs answer "what is this worth?" and stop there. That number is then used to justify a position size chosen by enthusiasm. This skill inverts the order: it computes a downside floor first, and derives the maximum position from the worst case rather than from how large the upside looks.

### 下行地板（worst-case）
- 假设：增长下修 40%、WACC +120bp、利润率/倍数压缩 0%
- **下行地板每股：25.33**
- 当前市价：18.00 ｜ 安全边际（地板/市价）：1.41×

### 仓位倒推（由最坏情况，非机会大小）
- **档位：小仓** ｜ 下行地板/市价=1.41×，仅小仓（≤组合 5–10%）

Quick start

python3 scripts/dcf_calc_test.py                        # verify the math (CI runs this too)
python3 scripts/dcf_calc.py --input examples/input.json # run the sample

Minimal input:

{
  "shares": 800, "net_debt": 1200, "price": 18.0,
  "base_fcf": 300, "years": 5,
  "growth": [0.15, 0.12, 0.10, 0.08, 0.06],
  "wacc": 0.09,
  "terminal": {"method": "gordon", "g": 0.025},
  "downside": {"growth_haircut": 0.4, "wacc_uplift_bp": 120, "margin_compression": 0.0}
}

Point your agent at SKILL.md. Stdlib only, no pip install.

What comes out

1. Base-case intrinsic value per share — 2-stage FCF, explicit period + terminal value
2. Sensitivity matrix — WACC × terminal growth, so you see how much of your answer is assumption
3. Downside floor — growth haircut, WACC uplift, margin compression applied together
4. Position cap — a tier, not a suggestion

Position sizing rule

Downside floor / market price	Tier	Max position
≥ 1.5×	Build	up to your normal portfolio limit
1.0× – 1.5×	Small	≤ 5–10% of portfolio
< 1.0×	Don't touch	0%

The floor is not a price target. It is the number that decides size.

Guardrails built into the script

• WACC ≤ g raises an error. A Gordon terminal value with growth at or above the discount rate is not a valuation, it's a typo with a decimal point.
• The downside floor must sit below the base case. If your stress case is higher than your base case, your assumptions contradict each other.
• Assumption ranges must be defensible. references/assumptions_bounds.md gives the ranges for WACC and growth; stepping outside them requires written justification.
• A >20% gap vs. comps blocks the conclusion.
• Methodology, including why terminal value dominates the answer and what to do about it: references/dcf_methodology.md

The 8 steps

Frame the target → get a clean FCF base → set two stages → pin the WACC → discount → run the sensitivity grid → stress to the floor → size the position.

Part of a three-skill loop

Skill	Question it answers
macro-dashboard	Should I be deploying capital at all right now?
valuation-comps	Is this cheap or expensive relative to its peers?
dcf-quick (here)	What is it worth, what is my downside, how big a position?

Check the macro switch before you size anything. A great floor in a risk-off regime is still a bad trade.

License

MIT. A model is not a forecast. The decision stays yours.