The Receipt Always Comes

In March 2023, a trader I'll call Mike (he posted his wallet on Reddit, so it's public record) deposited $50,000 into a Curve stablecoin pool advertising 12% APY. Eight months later, his portfolio showed $54,000. His actual net gain after gas fees, token incentives that dropped 40%, and the opportunity cost of not holding USDC in a high-yield savings account: roughly $1,200. Or about 2.4% annualized.

Curve's front page showed 12%. Mike got 2.4%.

This isn't a failure of DeFi. It's a failure to read the math.

Yield farming returns are the most misunderstood numbers in crypto. They're quoted, compared, and chased without any understanding of what's actually being measured. This piece is about closing that gap. By the end, you'll know how to evaluate a yield opportunity the way someone who actually does this professionally does—not by the headline number, but by what's left after the math settles.

What APY Actually Measures (And What It Doesn't)

Annual Percentage Yield is a rate of return on your capital, expressed as a yearly figure. Simple enough. The problem is that in DeFi, "yield" comes from at least five distinct sources, and most farmers never separate them.

1. Trading fees — Real income from people actually using the pool. On a Uniswap ETH/USDC pool during volatile periods, this might be 0.3% of volume flowing through. During quiet markets, near zero.

2. Token incentives — Project-native tokens paid to liquidity providers. These are often 70-90% of the advertised yield. And they come with a lockup, a vesting schedule, or a price that bleeds out before you can sell.

3. Borrow/lend spreads — On Aave or Compound, yield comes from interest borrowers pay. This is real, sustainable, and often modest—3-8% for stablecoins in normal conditions.

4. Rebase tokens — Some protocols like Terra's old anchor protocol paid yield by minting new tokens. The APY was real in nominal terms. The purchasing power was not.

5. Leverage amplification — Using borrowed assets to increase position size, multiplying both gains and losses. Usually the yield hunter's graveyard.

When you see "45% APY on staked ETH with partner tokens," you're probably looking at something like: 3% real yield + 42% in volatile project tokens that will be worth 60% less in three months.

The first step to evaluating any yield farm is asking: "What percentage of this number is trading fees and borrow interest?" If the answer is under 10%, you're farming tokens, not yield.

The Impermanent Loss Equation (Actually Explained)

Impermanent loss is the most important concept in liquidity provision, and also the most consistently misunderstood. Here's the actual math.

When you deposit assets into an AMM (automated market maker) like Uniswap or SushiSwap, you're providing liquidity in a pair—say, ETH and USDC. The AMM maintains a constant product formula: x × y = k. When ETH rises, the pool sells ETH into it, capping your upside. When ETH falls, the pool buys ETH, cushioning your loss. But the math isn't symmetric.

The formula for impermanent loss (IL) is:

IL = 2√r / (1 + r) - 1

Where r = current price / entry price

Let's run it with real numbers. You deposit when ETH is $2,000. You put in $1,000 ETH and $1,000 USDC. ETH doubles to $4,000.

r = 4,000 / 2,000 = 2

IL = 2√2 / (1 + 2) - 1 IL = 2(1.414) / 3 - 1 IL = 2.828 / 3 - 1 IL = 0.943 - 1 IL = -5.7%

Your impermanent loss is 5.7% of your position value. Meanwhile, the HODL comparison (just holding the ETH and USDC) would have returned 50% on the ETH portion. You've gained 44.3% instead of 50%—and that's your IL cost.

At ETH = $8,000 (4x), IL jumps to 25%. At ETH = $1,000 (0.5x), IL is -5.7% again (symmetric in percentage terms, but you're now holding more ETH than you started with).

The trap: impermanent loss becomes permanent the moment you withdraw. Until then, it's just "unrealized." But farmers often get forced to withdraw during volatility—liquidation risk, needing liquidity elsewhere, or panic. When IL becomes real, it hits all at once, often after you've been bragging about "earning 45%."

Reading a Real Farm: Aave USDC

Let's take a protocol with actual staying power and do the real math.

Aave V3 USDC lending currently yields approximately 5.2% APY (as of recent market conditions). Here's what that actually means for $10,000:

  • Gross yield: $520/year, or about $43/month
  • Network fees: At $5 gas per transaction, weekly compounding costs $260/year
  • Effective yield with weekly compounding: roughly 2.6%

If you compound daily on mainnet, gas might run $50 per transaction. That's $18,250/year in gas. You'd be net negative. This is why most retail yield farmers lose money on small positions.

The numbers only work at scale, or on L2s where gas is cents not dollars.

On Arbitrum, daily compounding on Aave costs maybe $2 in gas per year. Now your $10,000 earns $518 minus $2, yielding $516—essentially the advertised rate.

Lesson: For stablecoin lending under $25,000, stick to Arbitrum, Optimism, or Base. For anything under $5,000 in stablecoins, the gas math on any chain makes it barely worth the complexity.

The Compounding Trap: When 800% APY = 0% Real Return

Token incentives create absurd APY numbers that look like printing money. Here's why they're usually not.

A new protocol launches. They want liquidity. They offer 500% APY in their governance token for stablecoin LPs. You deposit $10,000.

After a month, you've earned $4,167 in their tokens. Great, right?

But their token launched at $2. It's now $0.80. You've earned $4,167 × $0.80 = $3,333 in nominal value. The token drops another 30% before you can sell (lockup period). You get $2,333. Gas and swap fees cost another $200. Your net: $2,133 on a $10,000 deposit over 30 days.

That's 21.3% annualized in token terms, but you locked in a 79% loss on the token itself. The APY was real. The purchasing power was not.

This is the standard business model of new DeFi protocols. Attract yield farmers with inflated APYs. Let token value decay. Farmers chase the number, not the value behind it.

How do you evaluate token incentives properly?

  1. Calculate your cost basis in the token. What will you receive, when can you sell, and at what price will that sale happen?
  2. Get the unlock schedule. Most token incentives vest over 6-12 months with weekly or monthly unlocks. Budget accordingly.
  3. Check the fully diluted valuation. If the token market cap is $5M but fully diluted is $500M, you're holding something that will be diluted by 99% over the next few years. Your "earnings" will be worthless.
  4. Assess sustainability. If 80% of the APY is token incentives, assume those tokens will drop 70% minimum. Adjust your expectations accordingly.

The Strategies That Actually Work

After watching thousands of farmers get rekt by the math, the strategies that consistently work come down to a few principles:

1. Stablecoin lending on L2s. Aave, Morpho, or Silo on Arbitrum or Base. For $10K-$500K, you're earning 4-6% with near-zero IL risk. Not exciting. Not sexy. Actually profitable.

2. Real yield protocols with sustainable revenue. Curve's veCRV model, Convex Finance, and similar protocols that distribute actual trading fees—not freshly minted tokens. The APY is lower (often 5-15% in CRV/CVX terms), but it's income, not dilution. When CRV spikes, you win. When it dumps, you're still earning.

3. Concentrated liquidity for professionals. On Uniswap V3, if you have a view on price range, you can earn 5-10x the normal trading fee income. But this requires active management, correct directional assumptions, and willingness to accept full IL if wrong. Most people should not do this.

4. ETH staking with leverage (for sophisticated traders). Using Lido stETH or Rocket Pool ETH as collateral to borrow stablecoins, then lending those stablecoins to earn the spread. At current rates, you can generate 8-12% on ETH-collateralized positions with managed liquidation risk. This is not for beginners.

The Mistakes That Kill Returns

Chasing headline APY. A 200% APY farm might be returning -50% in real terms after token decay. Always decompose.

Ignoring gas. On mainnet, daily compounding anything under $50K is likely a net negative in gas fees. Use L2s or accept lower frequency compounding.

Auto-compounding into the same token. If you're earning more of the LP token and it's inflationary, you're diluting yourself. You need to swap back into your original assets or you're just stacking tokens that keep expanding in supply.

Not tracking your actual entry. Most farmers don't calculate their blended entry price on a liquidity position. You might be earning fees but getting absolutely destroyed on IL. Run the numbers monthly.

Over-concentrating. Putting 50% of your portfolio into a single yield farm is not diversification. It's a levered bet on that protocol's token and smart contract security. Spread across 3-4 protocols with different risk profiles.

The Risk-Adjusted Framework

Here's how to evaluate any yield opportunity:

Sustainable yield = Trading fees + Borrow/lend spreads + Real protocol revenue Speculative yield = Token incentives with active markets, clear unlock schedules, and tokens you actually want to hold

Treat speculative yield as an option on the token's future value, not guaranteed income. Size accordingly—never more than you can afford to lose 100% of if the token goes to zero.

For most people, the answer is: stablecoin lending on Aave/Morpho at 4-6% on Arbitrum. Not exciting. Actually profitable. No IL. No token decay. No complex position management.

The yield farming industry wants you to believe complexity equals alpha. Usually, it just equals fees and IL dressed up in numbers you're not qualified to interpret.


The Real Takeaway

The math of yield farming isn't complicated—but the industry's marketing is designed to obscure it.

Before entering any farm:

  1. Calculate what percentage of APY is trading fees vs. token incentives. Under 15% fees means you're speculating on a token, not earning yield.
  2. Run the IL formula for your entry price. Know your break-even before you start.
  3. Price out gas for your compounding frequency. Daily on mainnet only works at $100K+. Otherwise, use L2s.
  4. Size token incentive positions as speculation, not income. Never more than you can lose entirely.
  5. Track your real returns quarterly. Not the number on the dashboard—the actual USD value of your position minus entry costs.

The farmers who make money aren't the ones chasing the highest APY. They're the ones who understand what the APY actually measures—and position accordingly.

Most of you are Mike. Don't be Mike.

---TITLE--- The Yield Farming Truth Table: What the APY Numbers Actually Mean When You Run the Numbers

---EXCERPT--- Most yield farmers are losing money while believing they're earning it. The advertised APY is a mirage built on compounding tokens you don't want, volatile assets you didn't account for, and a fee structure nobody reads. Here's the math that matters.

---META--- The real math behind DeFi yield farming: how to calculate actual returns and avoid the traps that make 80% of farmers net losers.

---TAGS--- yield farming, DeFi, APY, impermanent loss, liquidity pools, DeFi returns, crypto yield, DeFi math