Experiment 20260621-BROD-a

The effect of different rewiring strategies on the number of opinion/belief echo chambers differs from what has been reported in [20260621-BROD]; hence correcting the bug in the previous notebook does not allow to provide the same conclusion.

Experiment design

Date: 2026-08-26

Designer: Hiro KATAOKA (University of Tsukuba)

Hypotheses: The overall results are preserved with corrected implementation of the measures

100 agents; 2000 games

Variables

fixed variables: WORKFLOW TOPIC ATOM PACTIVE PREWRITE NBITERATIONS NBAGENTS NBRUNS ALPHA MU VALUE PREHOC

controlled variables: EPSILONS DELTAS SEEDS REWIRINGS

dependent variables: eo eb po pb uo ub

Values

EPSILONS: ['0.05', '0.1', '0.15', '0.2', '0.25', '0.3', '0.35', '0.4', '0.45', '0.5']
DELTAS: ['1', '2', '3', '4', '5', '6', '7']
SEEDS: ['544371', '315590', '903997', '779659', '556198', '160590', '103848', '94821', '501765', '722393', '908149', '490361', '170743', '808653', '204557', '720126', '24083', '756485', '970547', '909432']
REWIRINGS: ['minbo2', 'minob2', 'norewiring', 'swapMinMaxNbo', 'swapMinMaxNob', 'maxnbo', 'randomWCbo', 'maxnob', 'randomWCob', 'maxob', 'maxbo']

Measures

Before testing the hypothesis, we define the measures needed to check them. Let $A$ be the set of agents. Let $\mathcal S^t$ be the set of partitions of the network of agents at time $t$.

eo (Opinion echo chamber)

This measure counts the number of partitions such that:

  • they are segregated, i.e., $L^t(C)\leq\theta$ where $0<\theta<1$;
  • opinions are homogeneous, i.e., $M_O^t(C)\leq 10^{-4}$;
  • opinions have been reinforced, i.e., $D_O^t(C)$

where

$$ L^t(C)=\frac{|\{(a,a')\in N^t;a\in C\land a'\notin C\}|}{|\{(a,a')\in N^t;a\in C\}|}, $$

$$ M_O^t(C) = \max_{a,a'\in C}|O_a^t-O_{a'}^t| $$

and

$$ D_O^t(C)=\forall s\in [t_C,T),M_O^s(C)\geq M_O^{s+1}(C). $$

Here, $[t_C,T]$ is the maximal time window such that $\forall t\in[t_C,T]$, $C\in\mathcal S^t$ and $T$ is the number of iterations.

More formally, $eo$ is defined as:

$$ eo^t=|\{C\in\mathcal S^t;L^t(C)\leq\theta\land M_O^t(C)\leq 10^{-4}\land D_O^t(C)\}| $$

eb (Belief echo chamber)

This measure counts the number of partitions such that:

  • they are segregated, i.e., $L^t(C)\leq\theta$ where $0<\theta<1$;
  • beliefs are homogeneous, i.e., $M_B^t(C)=0$;
  • opinions have been reinforced, i.e., $D_B^t(C)$

where

$$ M_B^t(C) = \max_{a,a'\in C}d_B(B_a^t,B_{a'}^t) $$

and

$$ D_B^t(C)=\forall s\in [t_C,T),M_B^s(C)\geq M_B^{s+1}(C). $$

Here, $d_B$ is the Hamming distance over the models of two beliefs:

$$ d_B(B,B')=|\mathcal M(B)\setminus\mathcal M(B')|+|\mathcal M(B')\setminus\mathcal M(B)|. $$

More formally, $eb$ is defined as:

$$ eb=|\{C\in\mathcal S^t;L^t(C)\leq\theta\land M_B^t(C)=0\land D_B^t(C)\}| $$

po/pb (Population in opinion/belief echo chambers)

These measures count how many agents are in opinion/belief echo chambers, respectively.

uo (Unique opinions shared within opinion echo chambers)

This measure counts how many opinions are shared within opinion echo chambers. Let $A_O\subseteq A$ be the set of agents in opinion echo chambers. Then, this measure (roughly) returns $|\{O_a;a\in A_O\}|$.

To avoid applying strict equivalence check, i.e., $=$, toward several float numbers, we apply the hierarchical clustering with the allowed maximal distance $10^{-4}$.

ub (Unique belief shared within belief echo chambers)

This measure counts how many beliefs are shared within belief echo chambers. Let $A_B\subseteq A$ be the set of agents in belief echo chambers. Then, this measure returns $|\{B_a;a\in A_B\}|$.

Experiment

Date: 2026-08-26

Performer: Hiro KATAOKA (University of Tsukuba)

The whole experiment, from scratch, can be executed through:

In principle, this could be generated from command line through:

# This experiment is a reanalysis of existing one,
# hence no executions of the SOBA simuator are needed

# Python and library version are frozen by uv
$ uv sync

# The analysis is done through jupyter
$ bash utils/launch.sh &
# Do not forget to trust the notebook

# Before commiting the notebook
$ uv run nb-clean -e notebook.ipynb

# suppresses results and experiments
$ bash utils/cleanup.sh

# bash utils/anonymize.sh

Parameter file: params.sh

Hardware: AMD EPYC 7302P (16) @ 3.000GHz, Memory 128GB

OS: Ubuntu 22.04.5 LTS x86_64

Raw Results

This experiment is a reanalysis of the existing experiment [20260210-BROD], [20260610-BROD], [20260615-BROD], and [20260621-BROD]. Hence, no new simulations (or runs of the SOBA simulator) are needed.

To perform the following analysis, obtain the raw results of the experiment from these repositories and move all files under the directory called 'results'.

Note that files required to perform the analysis are downloaded by executing the script file utils/unzip.sh (be careful, this script automatically access to Zenodo. Please do not execute the file frequently).

Analysis

First, we compute the measures.

Note that the implementation of the measures has been corrected from [20260621-BROD].

['minbo2', 'minob2', 'norewiring', 'swapMinMaxNbo', 'swapMinMaxNob', 'maxnbo', 'randomWCbo', 'maxnob', 'randomWCob', 'maxob', 'maxbo']

Alternatively and to run the script stand alone, use notebook.py (which is generated from the notebook so far).

For the statistical tests, we will use 0.01 as the significance threshold.

Effect on opinion echo chambers

The average number of opinion echo chambers in each rewiring strategy is:

Out[12]:
eo
rewire
minbo2 0.000000
minob2 0.000000
norewiring 0.000000
swapMinMaxNbo 0.000000
swapMinMaxNob 0.000000
maxnbo 0.158571
randomWCbo 2.060000
maxnob 1.262143
randomWCob 2.651429
maxob 4.867857
maxbo 3.625000

This results show that: $$ \begin{aligned} \text{minbo2}\\ \text{minob2}\\ \text{no rewiring}\\ \text{swapMinMaxNbo} \\ \text{swapMinMaxNob} \\ \end{aligned} < \text{maxNbo} < \text{maxNob} < \text{randomWCbo} < \text{randomWCob} < \text{maxbo} < \text{maxob} $$

This is different from what has been obtained so far:

$$ \begin{aligned} \text{minbo2}\\ \text{minob2}\\ \text{no rewiring}\\ \text{swapMinMaxNbo} \\ \text{swapMinMaxNob} \\ \end{aligned} < \begin{aligned} \text{maxNbo}\\ \text{randomWCbo} \end{aligned} \leq \begin{aligned} \text{maxNob}\\ \text{randomWCob} \end{aligned} < \begin{aligned} \text{maxob} \\ \text{maxbo} \end{aligned} $$

Now we test ANOVA. It shows that changing the rewiring strategy has significant effect on the number of opinion echo chambers:

Out[13]:
df sum_sq mean_sq F PR(>F)
C(Q("rewire")) 10.0 42397.430390 4239.743039 1229.534536 0.0
Residual 15389.0 53065.126429 3.448250 NaN NaN

Hence, we perform the post-hoc test:

Out[15]:
group2 minbo2 minob2 norewiring swapMinMaxNbo swapMinMaxNob maxnbo maxnob randomWCbo randomWCob maxbo maxob
group1                      
minbo2 nan False False False False False True True True True True
minob2 False nan False False False False True True True True True
norewiring False False nan False False False True True True True True
swapMinMaxNbo False False False nan False False True True True True True
swapMinMaxNob False False False False nan False True True True True True
maxnbo False False False False False nan True True True True True
maxnob True True True True True True nan True True True True
randomWCbo True True True True True True True nan True True True
randomWCob True True True True True True True True nan True True
maxbo True True True True True True True True True nan True
maxob True True True True True True True True True True nan

This result supports the results from the comparisons of the average number of echo chambers.

Effect on belief echo chambers

The average number of belief echo chambers in each rewiring strategy is:

Out[16]:
eb
rewire
minbo2 0.107143
minob2 0.107143
norewiring 0.102143
swapMinMaxNbo 0.139286
swapMinMaxNob 0.136429
maxnbo 2.932143
randomWCbo 2.890000
maxnob 2.452857
randomWCob 2.657143
maxob 4.877857
maxbo 5.976429

This results suggest:

$$ \begin{aligned} \text{no rewire}\\ \text{minbo} \\ \text{minob} \\ \end{aligned} \leq \begin{aligned} \text{minNbo} \\ \text{minNob} \\ \end{aligned} < \begin{aligned} \text{randomWCbo} \\ \text{randomWCob} \\ \text{maxNbo} \\ \text{maxNob} \end{aligned} \leq \text{maxob} < \text{maxbo} $$

Compared to what is reported before (reposted below), minNbo and minNob can be classified in different groups from norewiring, minobo, and minob:

$$ \begin{aligned} \text{no rewire}\\ \text{minbo} \\ \text{minob} \\ \text{minNbo} \\ \text{minNob} \\ \end{aligned} < \begin{aligned} \text{randomWCbo} \\ \text{randomWCob} \\ \text{maxNbo} \\ \text{maxNob} \end{aligned} \leq \text{maxob} < \text{maxbo} $$

Now we perform ANOVA:

Out[17]:
df sum_sq mean_sq F PR(>F)
C(Q("rewire")) 10.0 61714.461169 6171.446117 1859.385421 0.0
Residual 15389.0 51077.298571 3.319078 NaN NaN

The results from ANOVA shows that the effect of changing the rewiring strategies on the number of belief echo chamber is significant.

Then we perform the Tukey HSD post hoc test:

Out[18]:
group2 minbo2 minob2 norewiring swapMinMaxNbo swapMinMaxNob maxnob randomWCob maxnbo randomWCbo maxbo maxob
group1                      
minbo2 nan False False False False True True True True True True
minob2 False nan False False False True True True True True True
norewiring False False nan False False True True True True True True
swapMinMaxNbo False False False nan False True True True True True True
swapMinMaxNob False False False False nan True True True True True True
maxnob True True True True True nan False True True True True
randomWCob True True True True True False nan True True True True
maxnbo True True True True True True True nan False True True
randomWCbo True True True True True True True False nan True True
maxbo True True True True True True True True True nan True
maxob True True True True True True True True True True nan

From this results:

  • we cannot see the significant difference between min{bo,ob}2 or norewiring and swapMinMaxN{bo,ob};
  • maxNbo/randomWCbo and maxNob/randomWCob should be classified as different groups.

Hence,

This results suggest:

$$ \begin{aligned} \text{no rewire}\\ \text{minbo} \\ \text{minob} \\ \text{minNbo} \\ \text{minNob} \\ \end{aligned} < \begin{aligned} \text{randomWCob} \\ \text{maxNob} \end{aligned} < \begin{aligned} \text{randomWCbo} \\ \text{maxNbo} \\ \end{aligned} < \text{maxob} < \text{maxbo} $$

Table 1

Out[21]:
    eps 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
delta kind measure                    
1 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 7.17 7.72 7.70 7.45 7.65 7.65 7.17 7.17 7.35 7.62
maximal2 eo 3.90 3.85 4.10 4.00 4.08 4.08 4.00 4.03 4.10 4.10
maximal eo 5.85 6.33 6.58 6.78 7.05 6.95 6.88 6.95 6.92 6.72
2 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 8.82 5.70 3.30 2.55 1.98 1.82 2.10 1.80 1.75 1.82
maximal2 eo 1.57 1.25 0.65 0.62 0.55 0.65 0.50 0.45 0.55 0.45
maximal eo 5.62 5.95 6.05 5.95 5.50 5.53 5.08 5.33 5.33 4.95
3 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 8.25 2.75 1.20 1.00 0.68 0.75 0.55 0.38 0.62 0.53
maximal2 eo 0.80 0.20 0.00 0.07 0.05 0.03 0.00 0.03 0.03 0.03
maximal eo 5.38 4.65 4.88 4.33 3.98 3.65 3.38 3.42 3.48 3.20
4 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 7.50 1.65 0.80 0.55 0.42 0.35 0.30 0.12 0.17 0.15
maximal2 eo 0.33 0.03 0.00 0.03 0.03 0.00 0.00 0.03 0.00 0.00
maximal eo 4.83 5.20 4.25 4.05 3.30 2.83 2.75 2.70 2.60 2.48
5 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 6.88 1.48 0.65 0.33 0.53 0.17 0.07 0.00 0.00 0.00
maximal2 eo 0.12 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
maximal eo 4.70 4.88 4.12 3.92 3.48 2.83 2.60 2.48 2.33 2.33
6 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 6.80 1.73 0.68 0.42 0.30 0.12 0.03 0.00 0.00 0.00
maximal2 eo 0.20 0.05 0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.00
maximal eo 4.35 4.67 4.30 3.88 3.10 2.65 2.38 2.75 2.55 2.12
7 no rewiring eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eo 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eo 7.03 1.38 0.50 0.42 0.23 0.12 0.00 0.00 0.00 0.00
maximal2 eo 0.15 0.05 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
maximal eo 4.47 4.95 4.20 3.83 3.05 3.05 2.67 2.42 2.38 2.25

Table 2

Out[23]:
    eps 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
delta kind measure                    
1 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eb 3.45 4.17 4.30 4.10 4.38 4.35 4.15 4.05 4.17 4.25
maximal2 eb 4.53 5.65 5.47 5.15 4.97 5.15 5.12 5.12 5.22 5.15
maximal eb 4.75 5.85 5.80 5.78 5.80 5.83 5.80 5.95 5.95 5.75
2 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eb 7.53 6.25 4.10 3.70 2.75 2.50 2.92 2.62 2.45 2.52
maximal2 eb 8.07 7.20 2.80 2.73 2.38 1.98 1.70 1.77 1.48 1.50
maximal eb 7.97 8.60 8.05 7.15 6.95 6.45 6.45 6.53 6.42 6.50
3 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
minimal2 eb 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
random eb 8.22 4.62 3.17 2.60 2.05 2.05 1.57 1.27 1.55 1.30
maximal2 eb 8.55 3.90 1.85 1.98 1.05 1.12 0.90 0.80 0.70 0.72
maximal eb 9.10 8.22 6.47 5.78 5.25 5.10 4.35 4.72 4.72 4.35
4 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.15 0.15 0.25
minimal eb 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.15 0.15 0.25
minimal2 eb 0.00 0.00 0.00 0.00 0.03 0.07 0.15 0.57 0.42 0.50
random eb 7.60 3.60 2.95 2.70 1.65 1.40 1.20 0.97 0.95 1.00
maximal2 eb 8.22 3.20 2.23 2.12 1.32 1.05 0.80 0.68 0.70 0.45
maximal eb 9.12 7.28 5.97 5.58 4.45 3.85 3.60 3.52 3.30 3.20
5 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.25 0.40 0.35 0.50 0.60
minimal eb 0.00 0.00 0.00 0.00 0.00 0.25 0.45 0.35 0.53 0.65
minimal2 eb 0.00 0.00 0.00 0.00 0.05 0.35 0.53 0.53 0.60 0.57
random eb 7.12 3.58 2.85 2.55 1.62 0.88 0.85 0.62 0.65 0.53
maximal2 eb 8.05 3.58 2.15 2.30 1.40 0.80 0.68 0.45 0.60 0.53
maximal eb 9.30 7.62 5.65 5.17 4.35 3.65 3.30 3.27 3.05 3.12
6 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.15 0.45 0.55 0.50 0.50
minimal eb 0.00 0.00 0.00 0.00 0.00 0.15 0.45 0.57 0.50 0.50
minimal2 eb 0.00 0.00 0.00 0.00 0.07 0.28 0.53 0.60 0.53 0.55
random eb 7.22 3.75 2.92 2.20 1.38 0.88 0.68 0.62 0.57 0.62
maximal2 eb 8.12 3.95 2.35 2.42 1.35 0.93 0.72 0.65 0.57 0.55
maximal eb 9.25 7.45 5.50 4.88 4.05 3.52 3.27 3.48 3.08 2.83
7 no rewiring eb 0.00 0.00 0.00 0.00 0.00 0.20 0.40 0.55 0.55 0.60
minimal eb 0.00 0.00 0.00 0.00 0.00 0.20 0.40 0.65 0.60 0.65
minimal2 eb 0.00 0.00 0.00 0.00 0.05 0.38 0.50 0.60 0.62 0.57
random eb 7.42 3.65 2.77 2.40 1.23 0.75 0.75 0.75 0.55 0.60
maximal2 eb 8.10 3.73 2.12 2.35 1.20 0.78 0.62 0.72 0.55 0.65
maximal eb 9.45 7.40 5.45 4.85 4.12 3.67 3.50 3.33 2.98 3.08

Table 3

Out[24]:
eo eb po pb uo ub
no rewiring 0.00±0.00 0.10±0.19 0.00±0.00 9.96±18.65 0.00±0.00 0.10±0.19
minimal 0.00±0.00 0.11±0.20 0.00±0.00 10.45±19.80 0.00±0.00 0.11±0.20
minimal2 0.00±0.00 0.14±0.23 0.00±0.00 13.42±22.22 0.00±0.00 0.14±0.23
random 2.36±2.98 2.77±1.99 20.58±26.88 68.42±15.80 2.01±2.43 2.76±1.98
maximal2 0.71±1.39 2.69±2.36 5.48±10.38 50.92±15.69 0.60±1.14 2.59±2.24
maximal 4.25±1.47 5.43±1.84 64.47±9.36 69.93±12.95 3.38±1.18 4.90±1.88

The largest difference is that when agents follow no rewiring they do not produce opinion (and belief in principle) echo chambers.

Why not single large echo chambers in $\varepsilon=0.5$, $\delta=7$, and no rewiring?

$\varepsilon=0.5$ and $\delta=7$ correspond to very tolerant agents.

For example, in the run norewiring-0.5-7-544371, the measure to detect opinion echo chambers works as follows:

{79}
Violates seg 4 0
{13}
Violates seg 7 0
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99}
Violates rei at t=121 (2.220446049250313e-15) and t+1=122 (2.3314683517128287e-15)

It identifies no echo chambers because the largest component violates the reinforcement property by the small difference in the maximal distance between $t=121$ and $t=122$ ($2.2\times 10^{-15}$ and $2.3\times 10^{-15}$, respectively).

What happens if we continue testing the reinforcement even if the reinforcement is the only one reasons not to be an echo chamber? If the largest component showed above violates it several times, all of the information should be printed below:

{79}
Violates seg 4 0
{13}
Violates seg 7 0
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99}
Violates rei at t=121 (2.220446049250313e-15) and t+1=122 (2.3314683517128287e-15)
Violates rei at t=115 (1.4210854715202004e-14) and t+1=116 (1.432187701766452e-14)
Violates rei at t=111 (1.609823385706477e-14) and t+1=112 (1.6431300764452317e-14)
Violates rei at t=110 (1.5765166949677223e-14) and t+1=111 (1.609823385706477e-14)
Violates rei at t=108 (4.451994328746878e-14) and t+1=109 (4.496403249731884e-14)
Violates rei at t=106 (7.394085344003543e-14) and t+1=107 (7.482903185973555e-14)
Violates rei at t=105 (7.283063041541027e-14) and t+1=106 (7.394085344003543e-14)
Violates rei at t=98 (6.674660824046441e-13) and t+1=99 (6.845635169838715e-13)
Violates rei at t=94 (1.7612578062653483e-12) and t+1=95 (1.8179902028236938e-12)
Violates rei at t=91 (5.01565455834907e-12) and t+1=92 (5.119682455756447e-12)
Violates rei at t=89 (2.0102142173072934e-11) and t+1=90 (2.0242474363385554e-11)
Violates rei at t=88 (1.9907520076856144e-11) and t+1=89 (2.0102142173072934e-11)
Violates rei at t=86 (1.9586665622739474e-11) and t+1=87 (1.9907520076856144e-11)
Violates rei at t=84 (5.993661122971616e-11) and t+1=85 (6.034350796824128e-11)
Violates rei at t=83 (5.963340932169103e-11) and t+1=84 (5.993661122971616e-11)
Violates rei at t=80 (1.1660294951809647e-10) and t+1=81 (1.1796563725852138e-10)
Violates rei at t=76 (1.921964809525889e-10) and t+1=77 (2.012839894760532e-10)
Violates rei at t=71 (3.2532666560669554e-09) and t+1=72 (3.2819246209569997e-09)
Violates rei at t=70 (3.195034015135434e-09) and t+1=71 (3.2532666560669554e-09)
Violates rei at t=69 (3.1814963996623646e-09) and t+1=70 (3.195034015135434e-09)
Violates rei at t=65 (2.4175180857177736e-08) and t+1=66 (2.437892110407347e-08)
Violates rei at t=64 (2.3987931085933667e-08) and t+1=65 (2.4175180857177736e-08)
Violates rei at t=63 (2.3717768637965264e-08) and t+1=64 (2.3987931085933667e-08)
Violates rei at t=61 (2.676207122576102e-08) and t+1=62 (2.6993903556871146e-08)
Violates rei at t=52 (5.17905361152593e-07) and t+1=53 (5.224606381970531e-07)
Violates rei at t=51 (5.058056862239368e-07) and t+1=52 (5.17905361152593e-07)
Violates rei at t=49 (1.9853413176784684e-06) and t+1=50 (1.987803085801154e-06)
Violates rei at t=48 (1.97622017672483e-06) and t+1=49 (1.9853413176784684e-06)
Violates rei at t=47 (1.965038504025962e-06) and t+1=48 (1.97622017672483e-06)
Violates rei at t=46 (1.9203118131194685e-06) and t+1=47 (1.965038504025962e-06)
Violates rei at t=45 (1.919120749094283e-06) and t+1=46 (1.9203118131194685e-06)
Violates rei at t=42 (6.168218166680006e-06) and t+1=43 (6.280759281107606e-06)
Violates rei at t=41 (6.011873863620565e-06) and t+1=42 (6.168218166680006e-06)
Violates rei at t=38 (2.222387621864108e-05) and t+1=39 (2.27289467300551e-05)
Violates rei at t=35 (7.061345995584833e-05) and t+1=36 (7.140280737638616e-05)
Violates rei at t=34 (7.055732116223457e-05) and t+1=35 (7.061345995584833e-05)
Violates rei at t=33 (6.941993620290798e-05) and t+1=34 (7.055732116223457e-05)
Violates rei at t=31 (0.00010810724769683144) and t+1=32 (0.00010823826764883027)

It violates several times while the difference is not large. For example, the difference in the maximal distance between $t=31$ and $t=32$ is around $1.3\times 10^{-7}$.

Hence, it is interesting to see how testing the reinforcement property can be relaxed.

Conclusion

  • The effect of different rewiring strategies reported in [20260621-BROD] has not been reproduced when the bug in the previous analysis is corrected.
  • This means that correcting the bug in the previous notebook does not allow us to conclude the same thing (which is problematic).

This file can be retrieved from URL https://sake.re/20260621-BROD-a