In our last post we introduced the framework for analysing liquidation slippage in $wAVAX and began investigating the microstructure of trading on Avalanche’s primary DEX venue LGJ (formerly Trader Joe). This time, we turned to a question that often comes up when assessing new collateral assets:

For curators setting risk parameters, the presence of JIT liquidity or MEV is extremely important for modelling

• On some chains, opportunistic liquidity providers or arbitrageurs jump in to smooth price impact, changing the effective slippage profile.

• If that were the case on Avalanche, we might need to factor in centralised exchange (CEX) order flow or external arbitrage support.

If these effects are absent, the modelling is much more straightforward: what you see on-chain is what you get. However, the lack of these effects can have profound impacts on the vault capacity.

We began with a large dataset of 1.26M transactions from LFJ’s wAVAX/USDC pool. From this, we identified:

• 17,124 JIT-like events (≈1.4% of txs)

• 153,371 MEV-like events (≈12.1% of txs)

We then regressed signed transaction volumes against prior activity, looking for the tell-tale diagonal “structure” that appears when JIT or MEV strategies are active.

Result

• JIT: Spearman ρ ≈ 0.02, (R^2 ≈ 0.0004)

• MEV: Spearman ρ ≈ 0.10, (R^2 ≈ 0.015)

Both relationships are statistically detectable thanks to the huge sample size, but economically negligible. In other words, the scatter plots look like Gaussian noise — no exploitable structure.

To stress test further, we isolated only:

• Correct-side JIT deposits (250 txs)

• Same-direction MEV swaps (9,049 txs)

Here again the signal was weak:

• JIT: no significant correlation (ρ ≈ −0.05, p = 0.43)

• MEV: statistically significant ρ ≈ 0.33, but still small in economic terms R^2 ≈ 0.05

This seems to confirm that even in larger trades, Avalanche pools aren’t seeing consistent JIT or MEV arbitrage flows.

There is no evidence of traders moving capital from centralised exchanges “bailing out” large trades with arbitrage support, rather it seems to suggest that trading is done on a slower-moving inventory rebalancing basis. To confirm this we would ned CEX data and calculate the autocorrelation of bin movements with respect to price but that is beyond the point of this analysis.

From a curator’s perspective, this is helpful: models of slippage and liquidation risk can be built directly from observed on-chain liquidity.

We don’t need to supplement our analysis with Binance or other CEX data when setting vault parameters.

Lack of JIT also presents a slight complexity because the absence of JIT means we should possibly separately consider other AMMs like Blackhole & Dodo as additive liquidity sources rather than assuming they will have commingled impact.

Their liquidity is non-trivial and can move liquidation costs down, but for now LFJ's exchange mechanics still dominate at scale due to the quantisation of the bin mechanics so we will continue to primarily focus on LFJ.

The absence of JIT/MEV is an opportunity for trading - both JIT and MeV are particularly crowded on Ethereum mainnet.

Yet, as of 24th August 2025 AVAX/USDC trades just under $1bn 24h total volume and >$70m on LFJ alone. meaning the CEX-DEX arbitrage opportunity is approximately half of that of Pendle's total x-chain volume on this single pool.

Maybe there is in an error in the data, as it's unclear why traders are fighting so hard over mainnet and ignoring such an opportunity.