五个月内10M收入流失：MEV对曲线的影响

olumeofboththevictimandtheattacker.WeneededtofilterouttheirvolumesinordertogetthevolumeforCurveliquiditypools.Inaddition,wenoticedthatthevolumeforsandwichbotstransactionswasnotasdistinctiveasaforarbitragebots,duetothespreadoutnatureoftheattacksthatwerenotdirectlytargetingtheCurveliquiditypools.Therefore,weneededtoconductamoreintensivetransactionanalysisinordertoobtainthetruevolumeforbothMEVtypes.

Thus,wedefinedthreevolumescalculations:

Sandwichvolume:Onlycountedthevictimsvolumethatgotcaughtinthesandwichattack

Arbitragevolume:CalculatedthevolumeforCurveliquiditypoolsonlyandexcludedvolumefromothervenues

Totalvolumeforarbitrageandsandwichbotstransactions:calculatedthevolumeforbothMEVtypesfromarbitragevolumeandsandwichvolume.

Figure7showsthetotalvolumecontributionfromarbitrageandsandwichbotstransactionsduringtheperiodfromJune1sttoOct31st.

Figure7:Totalvolumefromarbitrageandsandwichbotstransactions

Wecouldnoticeagappyphenomenoninthevolumeplot—thetotalvolumefellsharplyinJuly,withrecoveredgrowthinAugustandSeptember.

ArbitragevolumeandvolumefromtotalMEVbotstransactionshaddifferentvolumetrends.InJuly,thearbitragevolumeexhibitedhigherdrop,whichindicatesthatarbitragebotswerelessactivebecauseoftheoverallmarketconditions,suchashighertransactionfeesevaluatedbyEthereumgasprices.ThetotalvolumeofMEVtransactionsweremorestablecomparedtothearbitragevolumedespiteinthefallingtrend.

Asabovementioned,MEVbotstransactinthetradingvolumecompositely,influencingthetotalvolumeevenwhenitwasajointcontributionofdifferentbehavioralpatterns.WefurtherexpandedthisinvestigationintotemporaldistributionandvolumefrequencyanalysisforbothMEVtypes.

Conclusions

CurveisadeFiinfrastructureoperatingforstablecoinmetricswithuniquefeaturesandsupportedbythreetypesofliquiditypools.The3-tokenpoolanduniquepricecurvecanononlyreducefinancialrisksinherentinstablecoinsbutalsodeliverlesservicetocustomerswithlowerfeesandslippage.CurvedoesfaceMEVattacks,buthasgeneratedsignificantrevenuefrombotinterference.WeobservedequalinterestintheMEVmarketfrombotharbitrageandsandwichagents.Afterfilteringoutnon-Curvevolumedata,weobservedthatbotswerelessactiveinJulyduetohighEthereumgaspricesandoverallbearishmarketconditions.However,onestableobservationisbothMEVtypesgeneratedsignificantrevenuecomparedtothetotalrevenuegeneratedbyCurvepoolsthemselves.

Weshowedthroughouranalysisofthetotalvolumefromarbitrageandsandwichbotstransactionsthatarbitragetakesthemajorityofthevolumeduetothefactthatakeyparameterformarketfluctuations—pricechange—isoneofthefoundationalfeatureofarbitrageprofitopportunities.Sandwichattacks,however,arespreadoutsituationsanddonotdirectlyinflictthedecisivemarketimpact.Inmostcases,theytargetuserswithlessexpressivenaturewhencomparedtoarbitragebots.

Fromthevolumefrequencyanalysis,thesandwichattacksarefoundtohaveahighervolumepertransaction.Butduetothespreadoutnatureofthesandwichattackswecouldnoticeadecliningsandwichpercentageinthesummedtransfersofthatmeantotaltransactionevents.Basedontheseobservations,resilientdesignagainstbothMEVtypesisnecessary.Formoredetails,pleasereferencethemainbodyofthereport.

Overall,wehopeourreportwillprovidevaluableinsightsforstakeholdersandresearchersinterms ofthestrategiesagainstMEVactivitiesandhowwecanimprovetheefficiencyofpricediscoveryinAMMmarkets.