On National Teacher Day, meet the 2024-25 Kenan Fellows
Rich gets richer-Bitcoin Network
1. Rich Gets Richer !
Even In Bitcoin
Network?
Announcement of Current Research State..
Zehady Abdullah Khan
Lab of Professor Hidetoshi Shimodaira
Bachelor 4th year,
Mathematical Science Course,
Department of Information and Computer Sciences
School Of Engineering Science,
Osaka University.
3/30/2015
4. Unit Transaction & Network
Edges
Input AddressOutput Address
ID 1
ID 2
….
ID n_I
ID 1
ID 2
….
ID n_O 3/30/2015
4
# Possible Edge = n_I * n_O per tx
• Edges are not unique for simplicity
5. Growth of Bitcoin Network
3/30/2015
245 USD/BTC
Novermber 5, 2013
* What were the prime
Factors for
Price Increase ?
Bitcoin Exchange Started
2 Phases: 1. Initial Phase 2. Trading Phase
8. Wealth Inequality : Gini Coefficient
3/30/2015
0 £ G £1
G =
0 Perfect equality
1 Complete Inequality
ì
í
ï
îï
For a population uniform on the values
of di
9. G in Bitcoin
3/30/2015
Initially Phase: G for in degree is high Users mostly stored bitc
Trading Phase: Gin almost converged with Gout
12. Bitcoin Balance Statistics
3/30/2015
1 BTC = 383USD
(Nov 12,
2013)
Min 1st Q Median 3rd Q Max
BTC 0.00 0.00 0 0.00 111,100
USD 0 0 0 0 42,560,000
Total # of nodes Nodes with Non-
Zero Balance
Nodes with
Zero Balance
13,085,528 1,621,222 (12 %) 11464306 (88%)
Total # of BTC Total Generated BTC Left
21,000,000 11,942,900 9057100
15. 3/30/2015
Degree Centrality: Eigenvector Centrality
3/30/2015
Which are the most important or central vertices in the network ?
xi : the centrality of the node i
Aij : the adjacency matrix
i
j
Aij = 1
xi
'
= Aij xj
j
å
x'
= Ax
Repeat….
x(t) = At
x(0) : centrality vector after t steps
x(0) = civi ; vi is the eigenvector of xi
i
å
x(t) = At
civi = cili
t
vi =
i
å l1
t
ci
li
l1
æ
è
ç
ö
ø
÷
t
vi ---> c1l1
t
v1 (t-->¥)
i
å
i
å
Ax = l1x
xi = l1
-1
Aij xj
j
å
17. Eigenvector Centrality in Bitcoin
Network
3/30/2015
Largest Eigen Value, l1 = 761.6418
ID 3247203 3247205 3247200 3247206 3247202 3247197
Centralit
y 1 0.083688264 0.054296171 0.050453962 0.047525502 0.032626146
ID 3247199 3247191 3247196 3247209 3247193 3247215
Centralit
y
0.0244427
27 0.020694083 0.017641267 0.017346914 0.015898656 0.014387567
ID 3247213 3247190 3247195 3247194 3247180 3247192
Centralit
y
0.0138252
13 0.013527306 0.012857685 0.01278417 0.011125401 0.010684955
…………………………. Total 13041891 centralities.
18. Degree Centrality: Katz Centrality
3/30/2015
xi =a Aij xj + b
j
å
; a,b are positive constants
Here a keeps the balance.
x =aAx+ b1 ; 1=(1,1,1,...)
a -> 0 , x =b1
x = (I-aA)-1
1
(I-aA)-1
diverges when
| A -a-1
I | = 0 => a-1
= l1
Therefore, a <
1
l1
Node A has eigen vector centrality 0
Unexpectedly Node B has eigen vector
centrality 0
19. Page Rank
3/30/2015
xi =a Aij
xj
kj
out
+ b
j
å 1
For any j , if kj
out
= 0 , then put kj
out
=1
In Matrix Term,
x =aAD-1
x + b1
x = (I -aAD-1
)-1
1= D(D -aA)-1
1
a < l, where l is the largest eigenvalue of AD-1
For undirected graph,
l = 1 with eigenvector v = (k1,k2,k3,....)
If a vertex with high Katz centrality points to a large number of other vertices, all
those vertices will gain high centrality !!!!
i
All other nodes gains high
centrality if the node i
has high Katz centrality
20. Betweenness Centrality
3/30/2015
Measures the extent to which a vertex
lies on paths between other vertices.
ni
st =
1 if vertex i lies between the shortest path from s to t
0 else
ì
í
ï
îï
ü
ý
ï
þï
xi = ni
st;
st
å
xi = ni
st
s¹t
å
xi =
ni
st
gst
; gst = number of shortest paths from s to t
s¹t
å
21. Closeness Centrality
3/30/2015
Measures the mean distance from a vertex to other vertices.
Mean shortest path from i,
li =
1
n
dij
j
å ; dij : length of the shortest path from i to j
li =
1
n -1
dij
j¹i
å
Closeness centrality Ci
Ci =
1
li
Redefining,
Ci =
1
n -1
1
dijj¹i
å
Mean shortest path of the network
l =
1
n2
dij =
ij
å
1
n
li
i
å
22. References
1. Do the rich get richer? An empirical analysis of the BitCoin transaction
network. MIT tech
Daniel Kondor,∗ Marton Posfai, Istvan Csabai, and Gabor
Vattay ,Department of Physics of Complex Systems,
Eotvos Lorand University, Hungary
2. Networks An Introduction
M.E.J Newman
3. Albert, R. and Barabasi, A.-L. (2002). Statistical mechanics of complex
networks. Reviews of modern physics 74 (1), 47
4. Quantitative Analysis of the Full Bitcoin Transaction Graph.
http://eprint.iacr.org/2012/584
Ron, D. and Shamir, A. (2012).
5. http://bitcoin.org/about.html
6. http://www.vo.elte.hu/bitcoin
3/30/2015
24. What’s Next ?
Detail Dynamics of transaction
Non-parametric estimation of preferential
attachment function.
Network Visualization of important Bitcoin entity
Extracting Interesting Bitcoin Phenomenon
Bitcoin price prediction.
3/30/2015
Hinweis der Redaktion
Figure 1. The growth of the BitCoin network. Num- ber of addresses with nonzero balance (green), addresses in participating in at least one transaction in one week intervals (red) and the exchange price of bitcoins in US dollars accord- ing to MtGox, the largest BitCoin exchange site (blue). The two black lines are exponential functions bounding the growth of the networks with characteristic times 188 and 386 days.
Figure 2. Evolution of the indegree distribution. Since the beginning of 2011, the shape of the distribution does not change significantly. The black line shows a fitted power-law for the final network; the exponent is 2.18.
Figure 3. Evolution of the outdegree distribution. The black line shows a fitted power-law for the the final network; the exponent is 2.06.
Lorenz curve is a graphical representation of the cumulative distribution function of the empirical probability distribution of wealth
After mid-2010, the degree correlation coefficient stays between −0.01 and −0.05, reaching a value of r ≈ −0.014 by 2013,
In the initial phase C is high, fluctuating around 0.15 (see Fig. 5), possibly a result of transactions taking place between addresses belonging to a few enthusiasts try- ing out the BitCoin system by moving money between their addresses. In the trading phase, the clustering co- efficient reaches a stationary value around C ≈ 0.05 which is still higher than the clustering coefficient for ran- dom networks with the same degree sequence (Crand ≈ 0.0037(9))
We provide evidence that preferential attachment is an important factor shaping these distributions. Preferen- tial attachment is often referred to as the “rich get richer” scheme, meaning that hubs grow faster than low degree nodes. In the case of BitCoin this is more than an anal- ogy: we find that the wealth of rich users increases faster than the wealth of users with low balance; furthermore, we find positive correlation between the wealth and the degree of a node.