Atlas Asset Management · ZISIF §15

Conviction
Sizing

DISCOUNT × CONVICTION → TARGET SIZE
CONVEX, NOT LINEAR · ZERO AT FAIR VALUE
HARD-CAPPED AT POSITION LIMIT
SANITY CHECK · NOT INVESTMENT ADVICE
Conviction-based position sizer · interactive
Discount-to-intrinsic-value × conviction → target allocation. Zero at fair value. Convex, not linear.
Implied size
—
Room to add
—
Tranche size
—
Status
—
+25%
0.0%
10.0%
Allocation is zero at fair value, rises convexly with discount, hard-capped at the concentration limit.
Implied max size Hard cap Current position
How the model works
Conviction tiers
High convictioncap 11–14%
All three criteria present: durable moat, founder-led management, and a meaningful discount to intrinsic value (40%+). Reach full size earlier on the discount curve. Convex acceleration begins sooner.
Medium convictioncap 8–10%
Two of three criteria. Standard sizing with default convex curve. The most common case in the portfolio — most names sit here.
Low / speculativecap 5–7%
One or zero criteria. Smaller hard cap. Steeper discount required before convex acceleration kicks in. Used for opportunistic and special-situation positions.
Position layer
Core compounders
High-quality SMID businesses held for years. Larger caps allowed because thesis durability is high.
Opportunistic / EM
Geographic-edge plays in CEE, LATAM, Asia. Smaller hard caps reflect higher idiosyncratic risk per name.
Moonshot / event-driven
Binary catalysts. Sized small but real (2–6%). Hard cap kept low because outcome distribution is fat-tailed.
The discount-to-intrinsic-value input assumes you have done the underlying valuation work. The slider is the last 1% of arithmetic, not the analysis itself.

"Position size is a function of how wrong the market is, not of how excited we are. Conviction without discount earns no allocation — just observation."

— Atlas AM · sizing principle