Weighted-average anti-dilution

Market standard language (NVCA-style) does not reset the old conversion price all the way to the new cheap price. It blends them by how much money came in at the low price. That is weighted-average anti-dilution.

CP2 = CP1 × (A + B) / (A + C)

If you raised a lot at a slightly lower price, the adjustment is meaningful. If you raised a tiny amount at a fire-sale price, B is small relative to A, so CP2 barely moves — unlike full ratchet.

Walking through it

Before Series B: fully diluted A = 10,000,000. Series A conversion price CP1 = $2.00, on 2,000,000 preferred shares. Series B raises $2,000,000 at $1.00, so C = 2,000,000.

B   = 2,000,000 / 2.00 = 1,000,000
CP2 = 2.00 × (10,000,000 + 1,000,000) / (10,000,000 + 2,000,000)
    = 2.00 × 11 / 12
    ≈ $1.833

Conversion ratio = 2.00 / 1.833 ≈ 1.091. Series A's 2,000,000 preferred become about 2,182,000 common-equivalent — a bonus of ~182k shares.

Full ratchet on the same facts: CP2 = $1.00, ratio 2.0, Series A jumps to 4,000,000. More than 10× the weighted-average bonus.

Same setup, only a $200k down round at $1.00 (C = 200,000):

B   = 200,000 / 2.00 = 100,000
CP2 = 2.00 × (10,100,000) / (10,200,000) ≈ $1.980

Barely any adjustment. Full ratchet would still slam CP2 to $1.00. That is the fairness argument.

A larger down round, $8,000,000 at $1.00 (C = 8,000,000):

B   = 8,000,000 / 2.00 = 4,000,000
CP2 = 2.00 × 14 / 18 ≈ $1.556

Ratio ≈ 1.286. Series A becomes about 2,572,000 as-converted. Post-round fully diluted is roughly 10M + 8M + extra ~572k ≈ 18.57M. Series A goes from a pre-round 20% story to about 13.8%, versus 11.1% with no protection. Painful new-money dilution, softened — not erased.

Narrow-based vs broad-based

The fight inside the formula is what counts in A:

• Broad-based: outstanding common, preferred as-converted, and usually outstanding options/warrants (sometimes the whole option pool). Larger A → smaller adjustment → more founder-friendly. This is standard on NVCA forms.
• Narrow-based: might count only common and preferred, excluding options. Smaller A → harsher adjustment.

Common mistakes

• Memorizing the letters without defining A. Counsel's definition of "outstanding" is the deal.
• Skipping the tiny-round case. That is where weighted average and ratchet diverge most.
• Forgetting this changes conversion ratio, not the preferred certificate count. Cap tables often keep "2,000,000 Series A" and hide the new price in a footnote.

Check yourself

In the $2M-at-$1 example, why is B = 1,000,000?

B is money ÷ old price: $2,000,000 / $2.00. It is how many shares that money would have bought at CP1.

Does a larger A help founders or investors?

Founders. Larger A shrinks the adjustment, so preferred gets fewer bonus shares.

Key takeaways

• Weighted average is CP2 = CP1 × (A+B)/(A+C) — work an example before you sign.
• Tiny cheap rounds barely move CP; large cheap rounds move it more, still far less than a ratchet.
• Broad vs narrow is a fight about A. Read the definition.

Further learning

NVCA Model Legal DocumentsModel Legal Documents The NVCA Model Legal Documents serve as the industry-embraced model documents to be used in venture capital financings. What's new: Most recently, the NVCA model documents were revised to reflect…National Venture Capital Association - NVCA -