The second swap went through on the first try, with the slippage left low and no mysterious “price impact” surprise waiting at confirmation. The first attempt had been less elegant: the transaction sat there, failed, and left me wondering whether the network or the swap itself was the problem.
It was neither. The setting was wrong for the trade I was making.
The choice was not “high or low”
At the time, there were three obvious options. Leave the default alone, raise slippage until the transaction succeeded, or keep it tight and accept that the swap might need another attempt. The middle option was tempting because it appeared to solve everything. It also quietly gave the market more room to execute at a worse rate.
The better question was what the token and the route required. For a liquid token pair, a small slippage tolerance made sense. The expected price movement between signing and execution should have been limited, especially on Arbitrum, where the transaction itself was not the slow part. For a thinly traded token, a tighter setting could simply make the transaction fail repeatedly. That was a liquidity problem, not a reason to keep increasing the tolerance.
The practical check took less than a minute. The wallet had to be on Arbitrum One, with enough ETH available for gas, and the output amount had to look sensible before confirming. Then I checked the quoted price impact. If that number was already uncomfortable, changing slippage was treating the symptom rather than the trade.
What I use now
For a normal, liquid swap, I start with the smallest sensible tolerance shown by the interface. I only increase it when the trade is genuinely failing and the price impact still looks acceptable. I do it in small steps, not by jumping straight to a setting that can absorb a large price change.
That distinction is the useful part: slippage is permission, not a performance boost. A higher number does not make the route better. It only makes a wider execution price acceptable.
When the question is simply where to make the trade and inspect those details, the relevant reference is an arbitrum swap. The setting still deserves a deliberate look before the final confirmation.
That small pause is what I missed on the first attempt. It has since become the only part of the process I refuse to rush.