What is skew farming and how exactly does it work? Perpetual Pools is a derivative contract that swaps collateral between the long and short sides of one pool – when the collateral is unequal between these sides, the pool is 'skewed' which results in the swap agreement creating polarised leverage. Those who exploit the benefits of this particular leverage are called basis farmers, or skew farmers.

Like basis farming in Perpetual Swaps, these farmers can earn consistent returns while they remain delta-neutral. We've put together this guide to walk you through how to generate excess risk-adjusted returns:

1. Strategy

Overview

To reiterate, when the pool is skewed, the swap agreement creates polarised leverage. Traders in an under-collateralised side of a pool make more than the intended leverage when they win, while traders in an over-collateralised side make less when they win.

It's important to note that collateral skew has no effect on losses. Regardless of the skew, each side always loses at the intended leverage.

Constructing a Delta-Neutral Position

Create a delta-neutral position with equal notional value of: (a) pool tokens from an under collateralised side; and (b) the underlying asset.

Your first step is to mint pool tokens on the pools.tracer.finance site. Determine the most profitable pools to farm by clicking on "Browse" and then sorting by "Effective Gain". This sorting function lists pool positions with the greatest effective gain to the least.

For example, suppose 3S-ETH (Tracer's 3p short ETH derivative) has an effective gain of 3.3 (representing 10% more collateral in the long pool). Anyone holding 3S-ETH is effectively paid interest to have a short position.

Following the above example, consider theoretically taking a $1000 position in the 3S-ETH side of the pool. If you don't want short exposure and just want to collect the 'interest', you should delta hedge (to neutral) by acquiring long ETH exposure.

Say the ETH price is $3000 and you need to acquire +1 ETH to cancel out your short position. If this ETH was from a spot market, the total capital requirement is $4000 (1000 for pool position + 3000 for spot position). If you were to use a perpetual swap instead, with 10x leverage to hedge, the total capital requirement would be $1300 (1000 + 300).

Profiting from Skew

Traders can farm skew from this delta-neutral position. At the upcoming value transfer, the underlying price will either appreciate or depreciate, causing a value transfer to the long or short side of the pool, respectively.

Price appreciates: If the ETH price goes up 0.1% in a single hour, the long position would gain $3 (0.1% * $3000) and the 3S-ETH would lose $3 (0.1% * 3p leverage * $1000); netting $0.

Price depreciates: If the price goes down 0.1% instead, the long position would lose $3 (0.1% * $3000) and the 3S-ETH would gain $3.30 (0.1% * 1.1 skew * 3p leverage * $1000); netting $0.30. If the ETH always had an hourly volatility of 0.1%, the holder of these 2 positions could expect (on average) to make $0.15 per hour on $1300 of capital.

Profiting from PnL Compression

Another unique aspect of Perpetual Pools, the Power Leverage formula (used to determine the collateral swap), serves to increase yield for skew farmers. This formula behaves virtually the same as traditional leverage for normal price movements – but continuously dampens effective leverage as changes in the underlying price become more drastic. We call this effect, *PnL compression.

PnL compression* reduces both the profit and the loss of the derivative. What this means for skew farmers is that on the rare occasion that the underlying price moves by more than around 1 percent in a single hour, the PnL compression effect starts to become noticeable.

Price appreciates: Suppose the hedged positions described earlier were to experience a 5% increase in the underlying price in a single hour. The long position would gain $150 (5% * 3000) and the short position would lose $143 ($1000 - $1000 * 0.95^3), netting $7 profit for the skew farmer.

Price depreciates: Suppose that the hedged position were to undergo a 4.76% decrease (this decrease has an equal likelihood as the 5% increase) in the underlying price in a single hour. The long position would lose $143 (4.76% * 3000) and the short position would gain $151 (1000 * 1.1 skew * (1 - 0.952^3)); a net gain of $8.

See that the returns (PnL) for the pool tokens are "compressed" using the Power Leverage formula and the PnL for the other position is greater as a result. This strategy also works without skew, which we'll explain in more detail in Section 3.

Returns from PnL Compression Alone

The chart below simulates a spot-hedged skew farmer's performance holding 3S-ETH + ETH spot with no skew and 0% rebalance slippage over 2.6 years. This should demonstrate the profits from PnL compression bias alone:

All movement in the chart is generated by PnL compression bias. Occasionally it does lead to abrupt losses during hours where the underlying price depreciates rapidly, but the sudden gains are greater.

Overall, the strategy generates yield even when the collateral in the pools is equal. You may notice the period between 22k hours to 23k hours is a period of higher yield. This part of the simulation shows the volatile bull run where ETH price went from $650 to $1700 in only 1000 hours. Likewise, the largest loss for the skew farmer (occurring at hour 15248) was when ETH price dropped 9% in a single hour.

This strategy yields when there's PnL compression; and this compression occurs most often in times of great volatility, especially volatility where the price trends upwards.

Expected Returns

Rebalancing the long portion of a delta-hedged position hourly (to cancel out short exposure) will incur a slippage and trading fee expense for the skew farmer – this slippage will impact yield.

The figure below represents a simulation of skew farming with a spot hedge. It outlines the yield experienced by a skew farmer for various slippages at different pool skew levels. It's reasonable to expect (though we can't guarantee) similar yields. Testing for periods of price depreciation or stasis shows that ETH price decline has a negligible or inconsequential effect on expected return.

The parameters of the simulation assume hourly rebalancing, 2.6 years of actual ETH price history (starting at $500 and ending at $1700), while holding 3S-ETH and 2.97 * ETH maintained in equal dollar value. No leverage is used on the ETH hedge – meaning 25.2% of capital is 3S-ETH and 74.8% of capital is ETH. The APYs associated with a 32% skew were so significant that they didn't fit within the cell.

Figure 1 - APY returns associated with spot farming

2. Strategy in Practice

Remaining Delta-Neutral

The yields above assume hourly trading around the clock, something that may require a programmed script or bot. It's possible to manually rebalance once per day to get similar gross yields, but net yield is higher because you'd have decreased your slippage expense by lowering the volume traded – but this does include intermittent partial exposure to the underlying price.

For those choosing to manually rebalance once per day, the steps are:

a) Each day, look at the dollar value of your 3S tokens and multiply this number by 3. This is the dollar value of long ETH exposure required to hedge your short exposure.

b) Divide this by the current price of ETH and adjust your ETH holdings on a perpetual swap or spot market to match this number.

Put another way, you're adjusting your underlying holdings to be 3 times the dollar value of your 3 leverage short holdings daily (or as often as you wish).

Utilising Perpetual Swaps

The yield shown in the chart can usually be increased by using leverage on the long position. It's much more capital efficient to farm this way. There are multiple ways to obtain a leveraged long position; you could use a lending protocol, bank loan, or perpetual swap.

Perpetual swaps have similar interest rates as the other options but offer higher leverage and lower transaction costs. For these reasons, perpetual swaps are the recommended route.

A leveraged position on a perpetual swap increases capital efficiency but comes with the cost of interest via funding rate. Perpetual swaps typically have a funding rate of 0.01% every 8 hours or 11% per year (8766 hours in year * 0.01 / 8 hours funding period).

Suppose a skew farmer expects to get 30% annual yield according to the chart above. They could reduce their capital requirements by using a 10x leverage long position on a perpetual swap. The capital required would be $1300 (1000 3S-ETH + 300 10xETHPERP) rather than $4000 and because yield is measured relative to capital, the yield would be 3.08x as great using this strategy (1/(1300/4000)). Before funding rate expenses, this takes yield from 30% to 92%.

The 11% per year expected funding rate applies to the full notional value of the perpetual swap position (which is 10 times higher than the margin used when on 10x leverage). The $300 in margin is thus losing 110% per year. This margin is only 23% (300/1300) of the capital used so, relative to total invested by the skew farmer, the perpetual swap position is costing 27.5% per year (25%*110%). Subtracting the perpetual swap funding rate cost from the gross yield gives a net yield of 64.5%. In summary, the yields shown in Figure 1 can be higher using leverage for the long position.

Balancer Pools

If you prefer to avoid using a bot or are discouraged by the daily effort it takes to manually skew farm, a simple solution is available. Find the 3 leverage markets with the highest positive rebalancing rate and deposit into the Balancer pool containing the asset. The Balancer pools with greater than 1 leverage are both delta neutral and skew farming. They're also being rebalanced back to delta neutral by arbitrage bots constantly.

Only half the capital in the Balancer pools is used to farm the skew, so yields will be half of what's shown in the yield chart. While AMM liquidity providers also get the benefit of collecting swap fees from traders, they can lose money through impermanent loss. The unpredictable effect of impermanent loss could exceed the yield from skew farming. Read our Guide to the Balancer pools for more information.

3. Extended Description of PnL Compression

Let's look again at the example given in Section 1's Profiting From PnL Compression:

Price appreciates: Suppose the hedged positions described earlier were to experience a 5% increase in the underlying price in a single hour. The long position would gain $150 (5% * 3000), and the short position would lose $143 ($1000 - $1000 * 0.95^3) netting $7 profit for the skew farmer.

Price depreciates: Suppose that the hedged position were to undergo a 4.76% decrease (this decrease has an equal likelihood as the 5% increase) in the underlying price in a single hour. The long position would lose $143 (4.76% * 3000) and the short position would gain $151 (1000 * 1.1 skew * (1 - 0.952^3)); a net gain of $8.

The second trade was only profitable because of skew. With equal collateral between sides (no skew), PnL compression alone would cause a $137 gain by the short position. That's a $6 net loss for the skew farmer.

Outside of overhead costs like slippage, the PnL compression effect of power leverage is the only reason a skew farmer risks losing money. However, when there's no skew, one outcome (price appreciates) results in a $7 profit and the other (price depreciates) a $6 loss. This $1 discrepancy in favour of the skew farmer illustrates why, even with equal collateral in the pools, skew farmers still generate yield overall, from PnL compression.

The rare exposure to the underlying as a result of PnL compression means that skew-farming can't be labelled as having no price risk, though with appropriate options hedging, you could negate this price risk.

The occasional PnL compression means that both 3S-ETH and 3L-ETH tokens actually have an average leverage of 2.97 (as determined by analysing 2.6 years of ETH price history). Accordingly, a more exact hedge should use 2.97 times the exposure instead of 3, despite the token being labeled a 3 leverage token.

Distribution of PnL Compression for different Leverages

If you are interested in farming the excess risk-adjusted returns opportunity that pool skew presents, but require more information to do so, join the Tracer Discord to get in contact with the team or community.

You can also find our Skew Farming Guide here.