Can On-Chain Liquidi y Posi ion Da a Explain P o ide Beha io
in Uniswap V3?
Di yassh ee
Oc obe 7, 2025
Abs ac
This a icle uses on-chain liquidi y posi ion da a in Uniswap V3 o explain liquidi y p o ide
(LP) beha io . Using da a ex ac ed om he Non ungiblePosi ionManage con ac , we analyze
min and bu n e en s o cha ac e ize how LPs alloca e and adjus capi al ac oss p ice anges,
ee ie s, and ime. By modeling how liquidi y p o ide s his o ically adjus posi ions a ound
ma ke p ices and ola ili y, his me hodology aims o p o ide a amewo k o using posi ion
da a as po en ial indica o s o ma ke dynamics. The p oposed me hod can be applied bo h
o his o ical da ase s and eal- ime da a s eams o moni o liquidi y posi ion anges, and he
playe s in he AMM space.
1 In oduc ion
Uniswap V3 imp o ed au oma ed ma ke make s (AMMs) logic by in oducing concen a ed liq-
uidi y, whe e liquidi y p o ide s can alloca e capi al wi hin speci ic p ice anges a he han ac oss
he en i e p ice spec um. This allows o mo e capi al-e icien liquidi y p o ision, bu in oduces
complexi y in unde s anding posi ion cha ac e is ics and his o ical liquidi y pa e ns.
The key ques ion his a icle add esses is: Can his o ical liquidi y posi ions e eal pa -
e ns ha help us unde s and liquidi y p o ide beha io and p edic p ice ac ion?
This a icle p esen s a comp ehensi e analysis o his o ical liquidi y posi ions ex ac ed om he
Uniswap V3 Non ungiblePosi ionManage con ac o in es iga e whe he posi ion pa e ns can
p o ide insigh s in o ma ke beha io .
Da a A ailabili y
All p ocessed da a and code a e a ailable in he Gi Hub eposi o y: h ps://gi hub.com/Di yn/
uniswap- 3-posi ion-analysis
1.1 Resea ch Objec i es
The p ima y objec i es o his esea ch a e:
1. Liquidi y P o ide Beha io Analysis: Unde s and whe e and when liquidi y p o ide s
choose o alloca e capi al
2. Ma ke Sen imen Indica o s: De e mine i posi ion concen a ion pa e ns can se e as
ma ke sen imen indica o s
3. P edic i e Value: Assess whe he his o ical posi ion da a can p edic u u e p ice ac ion
1
2 Me hodology
2.1 Da a Sou ce and A chi ec u e
Ou analysis u ilizes he Bi que y Uniswap APIs o que y on-chain da a di ec ly om he E he eum
blockchain. The me hodology ocuses on wo p ima y da a sou ces:
1. Posi ion Da a: Di ec calls o he posi ions unc ion o he Uniswap V3 Non ungiblePo-
si ionManage con ac (add ess: 0xC36442b4a4522E871399CD717aBDD847Ab11FE88)
2. Min E en s: His o ical min e en s ep esen ing liquidi y addi ion ansac ions
3. Posi ion C ea o s: Analysis o liquidi y p o ide s who c ea e posi ions h ough min e en s
The sys em a chi ec u e consis s o ou main componen s:
Lis ing 1: Bi que yClien Class S uc u e
1class Bi que yClien :
2de ge _his o ical_posi ions (sel , s a _da e , end_da e )
3de ge _ ecen _posi ions_ eal ime (sel )
4de ge _his o ical_min _e en s (sel , s a _da e , end_da e )
5de ge _ ecen _posi ion_c ea o s (sel )
6de ge _ oken_decimals (sel , oken_add esses )
2.2 Posi ion Da a Ex ac ion
The posi ion ex ac ion p ocess in ol es que ying he posi ions unc ion o each unique oken
ID. The logic o he que y is summa ized below:
Algo i hm 1 Re ie e Uniswap V3 Posi ions
Requi e: s a da e,end da e,con ac add ess =0xC36442b4a4522E871399CD717aBDD847Ab11FE88
1: Selec da ase = a chi e o eal ime, ne wo k = E he eum mainne
2: Fil e calls whe e Signa u e Name is "posi ions" and To equals con ac add ess
3: Res ic by Block Da e: a e s a da e and be o e end da e
4: Se pagina ion limi o X and o de by descending Block Numbe
5: Fo each call, ead: A gumen s ( okenId), Re u ns (posi ion ields), T ansac ion,Block
Ensu e: Ou pu lis o posi ions wi h ields: oken0, oken1, liquidi y, ee, ickLowe , ickUppe
Each posi ion e u ns he ollowing key pa ame e s:
• oken0 and oken1: The wo okens in he liquidi y pai
•liquidi y: The cu en liquidi y amoun (L)
• ee: The ee ie (500, 3000, o 10000 o 0.05%, 0.3%, o 1%)
• ickLowe and ickUppe : The p ice ange bounds in ick space
2
Algo i hm 2 P ice Band Calcula ion Algo i hm
Requi e: ick ∈Z, oken0decimals ∈N, oken1decimals ∈N
Ensu e: p ice ∈R
1: p ice unadjus ed ←(1.0001) ick
2: decimal adjus men ←10( oken0decimals− oken1decimals)
3: inal p ice ←p ice unadjus ed ×decimal adjus men
4: e u n inal p ice
2.3 P ice Band Calcula ion Algo i hm
The co e inno a ion o his me hodology is he p ecise calcula ion o p ice bands om ick da a.
Uniswap V3 uses a ick-based p icing sys em whe e each ick ep esen s a speci ic p ice poin .
This algo i hm accoun s o he exponen ial na u e o he ick sys em and adjus s o oken
decimal di e ences, ensu ing accu a e p ice calcula ions ac oss all oken pai s.
The ma hema ical ounda ion is based on Uniswap V3’s p icing o mula:
P= (1.0001) ick ×10(d0−d1)
Whe e:
•Pis he p ice o oken0 in e ms o oken1
• ick is he ick alue
•d0and d1a e he decimal places o oken0 and oken1 espec i ely
No e on Token Decimals: All on-chain da a equi es decimal no maliza ion, as ERC-20
okens s o e alues as in ege s scaled by 10decimals. Ou implemen a ion que ies oken me ada a o
ensu e accu a e con e sion o aw blockchain alues o human- eadable amoun s.
2.4 Min E en Analysis
To unde s and his o ical liquidi y addi ion pa e ns, we analyze min e en s ep esen ing new po-
si ion c ea ion:
Lis ing 2: Min A gumen s Pa se
1de pa se_min _bu n_a gumen s(a gumen s: lis ) -> dic :
2""" Pa se min / bu n unc ion a gumen s o ex ac posi ion pa ame e s ""
"
3pa ams = {}
4
5 o a g in a gumen s :
6index = a g. ge (’Index ’, -1)
7 alue = a g. ge (’Value ’, {})
8
9# Index mapping o min pa ame e s :
10 # 0: oken0 , 1: oken1 , 2: ee , 3: ickLowe , 4: ickUppe
11 # 5: amoun 0Desi ed , 6: amoun 0Min , 7: amoun 1Desi ed , 8:
amoun 1Min
12 # 9: ecipien , 10: deadline
13
3
14 i index == 0: pa ams [’ oken0’] = alue [’add ess’]
15 eli index == 1: pa ams [ ’ oken1’] = alue [ ’add ess’]
16 eli index == 3: pa ams [ ’ ickLowe ’] = in ( alue [ ’bigIn ege ’])
17 eli index == 4: pa ams [ ’ ickUppe ’] = in ( alue [ ’bigIn ege ’])
18 # ... addi ional pa ame e ex ac ion
19
20 e u n pa ams
Impo an Technical No e: In Uniswap V3’s Non ungiblePosi ionManage , min and bu n
e en s ope a e di e en ly wi h espec o posi ion iden i ica ion:
•Min e en s c ea e new liquidi y posi ions wi h ull pa ame e s ( oken0, oken1, ee ie ,
ick anges, amoun s, ecipien , deadline) and e u n a unique okenId ep esen ing he newly
min ed NFT posi ion.
•Bu n e en s des oy exis ing posi ions by e e encing only he okenId ( he NFT posi ion
ID), no he unde lying oken con ac add esses. This is because each posi ion is ep esen ed
as an ERC-721 NFT, and bu ning des oys he NFT i sel by i s ID a he han di ec ly
e e encing he oken pai .
This dis inc ion is c i ical o da a p ocessing: min e en s con ain comple e posi ion pa ame e s
in hei a gumen s, while bu n e en s p ima ily e e ence posi ions by hei NFT okenId.
2.5 Posi ion C ea o Analysis
To unde s and liquidi y p o ide beha io pa e ns, we analyze posi ion c ea o s h ough min
e en s o iden i y he mos ac i e and in luen ial liquidi y p o ide s. This analysis p o ides insigh s
in o he dis ibu ion o liquidi y p o ision ac i i y and helps iden i y key ma ke pa icipan s.
2.5.1 C ea o Da a Ex ac ion
The posi ion c ea o analysis ex ac s da a om min e en s o ack which add esses a e c ea ing
liquidi y posi ions:
Algo i hm 3 Re ie e Recen Posi ion C ea o s om Min E en s
Requi e: con ac add ess =0xC36442b4a4522E871399CD717aBDD847Ab11FE88,limi =X
1: Selec da ase = eal ime (o a chi e as applicable), ne wo k = E he eum mainne
2: Fil e calls whe e Signa u e Name is "min " and To equals con ac add ess
3: O de by descending Block Numbe and apply limi = limi
4: Fo each call, ead: T ansac ion.F om (c ea o ), T ansac ion.Hash,T ansac ion.Time,T ans-
ac ion.ValueInUSD
5: Op ionally pa se Re u ns alues (e.g., okenId o liquidi y pa ame e s) as needed
Ensu e: Ou pu lis o c ea o en ies wi h ac i i y and op ional alue me ics
2.5.2 C ea o Ranking Me hodology
The analysis anks posi ion c ea o s by mul iple me ics o unde s and di e en aspec s o liquidi y
p o ide beha io :
1. Posi ion Coun : To al numbe o posi ions c ea ed by each add ess
4
2. To al Liquidi y: Cumula i e liquidi y p o ided ac oss all posi ions
3. Unique T ading Pai s: Di e si y o oken pai s ac oss posi ions
4. Tempo al Pa e ns: Posi ion c ea ion iming and p ice band
2.5.3 C ea o S a is ics Calcula ion
Fo each c ea o add ess, we calcula e comp ehensi e s a is ics:
•Ac i i y Me ics: To al posi ions c ea ed, i s and las posi ion imes amps
•Liquidi y Me ics: To al liquidi y p o ided, a e age liquidi y pe posi ion
•Di e si y Me ics: Unique ading pai s coun , ee ie p e e ences
•Tempo al Me ics: Posi ion c ea ion equency, ac i i y pe iods
Figu e 1: Example o a Uniswap V3 liquidi y posi ion showing p ice band con igu a ion
2.6 Bu n E en Analysis
To complemen ou analysis o posi ion c ea ion h ough min e en s, we examine posi ion closu e
pa e ns h ough bu n e en s. As p e iously no ed, bu n e en s ope a e di e en ly om min
e en s: hey only e e ence he NFT okenId a he han con aining comple e posi ion pa ame e s.
2.6.1 Bu n E en Da a Ex ac ion
The bu n e en analysis ex ac s posi ion closu e da a o unde s and when and how liquidi y
p o ide s exi hei posi ions:
5
Algo i hm 4 Re ie e Bu n E en s
Requi e: s a da e,end da e,con ac add ess =0xC36442b4a4522E871399CD717aBDD847Ab11FE88,
limi =X
1: Selec da ase = a chi e (o eal ime as applicable), ne wo k = E he eum mainne
2: Fil e calls whe e Signa u e Name is "bu n" and To equals con ac add ess
3: Fil e by Block Da e be ween s a da e and end da e
4: O de by descending Block Numbe and apply limi = limi
5: Fo each call, ead: A gumen s. okenId,T ansac ion.F om (bu ne ), T ansac ion.Hash,
Block.Time
Ensu e: Ou pu lis o bu n e en s wi h okenId, bu ne add ess, and imes amp
2.6.2 Bu n Pa e n Analysis
Ou bu n e en analysis ocuses on se e al key me ics:
1. Token ID Bu n F equency: How many imes each NFT posi ion is bu ned
2. Bu ne Ac i i y: Dis ibu ion o bu n ope a ions ac oss walle add esses
3. Tempo al Pa e ns: Time-o -day and daily dis ibu ion o posi ion closu es
4. Bu ne Concen a ion: Iden i ica ion o high-ac i i y posi ion close s
The analysis e eals pa e ns in posi ion li ecycle managemen and can indica e ma ke sen i-
men shi s when co ela ed wi h min e en ac i i y.
3 Resul s and Analysis
3.1 Da ase Cha ac e is ics
Ou analysis p ocessed 5,711 unique liquidi y posi ions om he Uniswap V3 p o ocol, spanning
mul iple oken pai s and ee ie s. The da ase includes:
•His o ical Da a: Posi ion da a spanning om Sep embe 22 o Oc obe 7, 2025
•Real- ime Da a: Recen posi ions om he li e blockchain s a e
•Token Co e age: 205 unique ERC-20 okens wi h di e en decimal con igu a ions
•Fee Tie s: Analysis ac oss all Uniswap V3 ee ie s (0.01%, 0.05%, 0.3%, 1%)
No e on Time Window Selec ion: The choice o a 2-week maximum lookback window is
delibe a e and e lec s he dynamic na u e o decen alized exchange ma ke s. In DeFi, ma ke
condi ions, liquidi y p o ide s a egies, and oken pai dynamics e ol e apidly in esponse o
p ice mo emen s, p o ocol upda es, and b oade ma ke sen imen . His o ical da a beyond 2 weeks
o en becomes less ele an o unde s anding cu en liquidi y p o ision pa e ns and p edic ing
nea - e m beha io , as posi ion s a egies ha we e op imal weeks ago may no longe e lec cu en
ma ke eali ies.
6
3.2 Posi ion C ea o Analysis Resul s
Ou analysis o posi ion c ea o s e eals signi ican insigh s in o liquidi y p o ide beha io pa e ns.
F om he min e en s analysis, we iden i ied 3,541 unique posi ion c ea o s who c ea ed 11,315
o al posi ions du ing he analysis pe iod.
3.2.1 Top C ea o Pa e ns
The c ea o analysis e eals dis inc pa e ns in liquidi y p o ide beha io :
Mos Ac i e C ea o s:
•A e age Ac i i y: 3.20 posi ions pe c ea o , showing mode a e posi ion c ea ion equency
•Di e si ied C ea o s: Some c ea o s sp ead ac oss mul iple ading pai s
•Concen a ed C ea o s: O he s ocus on single oken pai s wi h mul iple posi ions
Liquidi y Concen a ion:
•High-Value C ea o s: Single ansac ions p o iding liquidi y wo h o e $5.16 million in
ansac ion alue, demons a ing signi ican ins i u ional pa icipa ion
•Fee Tie P e e ences: Posi ion c ea o s show s ong p e e ence o he 0.3% ee ie (39.5%
o min e en s), indica ing balanced ee/liquidi y en i onmen s, wi h signi ican ac i i y in
he 1% ie (28.6%) o mo e ola ile pai s
3.2.2 C ea o Beha io Insigh s
The analysis e eals se e al key pa e ns in c ea o beha io :
1. Ac i i y Clus e ing: Posi ion c ea ion ends o clus e a ound speci ic ime pe iods, indi-
ca ing coo dina ed o algo i hmic ading s a egies
2. Pai Specializa ion: Many c ea o s ocus on speci ic oken pai s a he han di e si ying
ac oss mul iple pai s
3. Scale S a egies: Di e en c ea o s employ di e en scales - om small, equen posi ions
o la ge, concen a ed posi ions
4. Tempo al Pa e ns: C ea o ac i i y co ela es wi h ma ke ola ili y, wi h inc eased po-
si ion c ea ion du ing pe iods o p ice unce ain y
7
Figu e 2: Min e en s by ee ie showing he dis ibu ion o e en s ac oss 0.01%, 0.05%, 0.3%, and
1% ie s
Figu e 3: Analysis o posi ion c ea o s showing dis ibu ion pa e ns and ac i i y me ics
3.3 Bu n E en Analysis Resul s
Ou analysis o bu n e en s o e a 2-week pe iod (Sep embe 24 - Oc obe 7, 2025) e eals signi ican
insigh s in o posi ion closu e beha io on Uniswap V3:
3.3.1 Bu n E en S a is ics
The da ase comp ises 2,197 o al bu n e en s execu ed by 344 unique bu ne add esses,
yielding an a e age o 6.39 bu ns pe bu ne . This concen a ion sugges s ha a ela i ely
8
small g oup o add esses a e esponsible o mos posi ion closu es, indica ing po en ial p o essional
liquidi y manage s o au oma ed ma ke -making s a egies.
Key Findings:
•Peak Ac i i y: Oc obe 6, 2025 saw 274 bu n e en s, ep esen ing 12.5% o all bu ns in
he analysis pe iod
•Hou ly Pa e n: Peak bu n ac i i y occu ed a 20:00 UTC wi h 32 bu ns
•Daily Dis ibu ion: Bu n e en s anged om 72 (Oc obe 7, pa ial day) o 274 (Oc obe
6), wi h an a e age o 157 bu ns pe day o e he 2-week pe iod
•Tempo al Clus e ing: Signi ican bu n ac i i y clus e ing indica es coo dina ed esponses
o ma ke e en s o p ice mo emen s
3.3.2 Token ID Bu n Pa e ns
Analysis o bu n equency by oken ID e eals in e es ing posi ion li ecycle pa e ns. Figu e 4
shows he dis ibu ion o bu n coun s ac oss NFT posi ions, whe e some posi ions a e bu ned
mul iple imes, indica ing ei he :
1. Posi ion ec ea ion cycles (min -bu n-min pa e ns)
2. Mul iple liquidi y adjus men s h ough comple e posi ion closu e
3. Posi ion mig a ion s a egies ac oss di e en p ice anges
Figu e 4: Token ID Bu n Coun dis ibu ion showing how many imes each NFT posi ion was
bu ned du ing he analysis pe iod
9
