Yet another dive in the New Zealand & NZWB data
Some progresses on the ongoing exchange with Steve Kirsch, and some answers to his questions.
Back to New-Zealand data, after direct exchanges with Steve Kirsch produced some unexpected benefits. My terms to consider pursuing the conversation were:
to retract a misleading article on the Pfizer trial deaths
and to edit an erroneous attribution of my previous response… which had been unexpectedly pinned on
.
These have been almost met during the latest Twitter back & forth (the article on Pfizer deaths was completely re-edited to fit the real figures, not retracted announcing the error rectification to his followers, as I had expected).
Contemplating, nevertheless, what were undeniable progresses, we (5shezz4 having the kindness to join me to ensure that my Frogglish was as clear as can be) first addressed the arguments left answered on Steve’s Substack, after my latest article.
We then explained the demographics at work on a new chart, for which Steve asked more explanations during our direct communications.
Latest arguments on Substack
The latest arguments advanced in Steve Kirsch’s own response on Substack were the following (in Italic, numbered, the arguments Steve had sustained - below in italic-bold his reply to the arguments - and lastly our own reply preceded by an arrow “→”).
Points this time include:
Steve then arguments with a chart showing an apparent 2022 rise in all cause mortality. No source. No code.
The data is available publicly. It is telling he doesn’t know the sources for mortality and vaccination data in New Zealand.
→ I reproduced the chart and refined the analysis. How do you do that without locating the data? Is this pure misrepresentation?
As anyone can see quite clearly on this second chart, the vaccinations occurred mostly after the spike of the 2021 deaths
Yes it did. I pointed out the the subsequent HUGE spike was AFTER the vaccine.
→ Yes, a year later, during the first COVID-19 wave, after an exceptional low in deaths which we highlighted and which you managed to completely ignore in your arguments summary. It was just.. well.. the main topic of the article.
The jabs (within the short term experience we currently have) will kill and maim mostly short-term due to variable-by-batches endotoxins impurities, what would be the suggested damage mechanism for deaths occurring around a year later, according to Steve ..?
That is not my concern. My concern is only on plausible mechanism. Injecting a toxic substance into people’s bodies can kill them.
→ Well, that’s annoying when one aims to have something acknowledged by the scientific community. Hard enough to have them acknowledge the obvious short term damages with known damage mechanisms and not a single analysis showing batches clean from endotoxins.
Long story short, Steve doesn’t understand the model.
Ad hominem attack. I’ve laid out the cohort time-series analysis and he NEVER finds ANYTHING wrong with ANY of my numbers. No error in buckets.py. No error in the v4 spreadsheet. No errors.
→ Of course I don’t find something in it, I haven’t felt the need to look at it so far, confer point 6.
The man simply doesn’t understand the notions of aging of a cohort or of seasonality of deaths.
Igor Chudov is reading my more extensive analysis (not yet published). He will vouch that OPENVAET statement is false and defamatory.
→ Igor would sure be a worthy opponent in a debate. I’ll respectfully wait that he takes position.
Steve can keep looking for excess deaths in a cohort and database under-registering deaths if it amuses him.
The figure above shows deaths are increasing. If the database is under-registering deaths as OPENVAET claims, then the figure above is even more damning. Where is the EVIDENCE for that? It’s only speculation. Is the database under-registering deaths at the end? At the start? By how much? And how do you know for sure? OPENVAET leaves us wondering.
→ Acknowledging this data isn’t under-reported in deaths would be a proof of vaccine efficacy, nothing else - with an overall mortality lower than expected compared to the NZ population over the same period, as we already demonstrated. It is clear on the provided chart when deaths are missing - and we provided an accurate estimate of the under-reporting factor.
Note: I offered Steve an alternative bet. Debate on the issue, the loser retires. Let’s hope he will take it.
and so on.
The bet is asymmetrical. People with pseudonyms can easily return. Also, I don’t believe in censoring anyone’s voice. Censorship is a bad thing. I won’t be a part of that.
→ We addressed this point during our exchanges - and our terms seem to have been accepted.
My response to all this is simple: can we stick to the NZ data? Here’s a chart of deaths per week of people who got shot 1 in July, Aug, Sept 2021 aged 80 and older. Simply explain how the deaths per week goes up by 50% or more over time. It’s supposed to be going down
→ The famous “It’s supposed to be going down” comes back. It comes from an erroneous assertion made during Steve Kirsch’s presentation at MIT that all the vaccines dosing schedules are the same, and that they are administered to the same populations - disregarding the fact that the COVID-19 ‘vaccines’ administrations have been quite specific.
While we already demonstrated that it “shouldn’t be going down”, and while Arkmedic already replied, we further illustrate below why the mortality in this “dose 1 cohort” doesn’t appear abnormal.
New Argument : Days from Dose 1 till Death
Moving on from the arguments listed on Substack we hadn't addressed ... At Steve's request via Twitter's Direct messages, we focused on demonstrating why we observe the following timing to deaths observed in the cohort.
Preliminary Analysis
Looking at this chart, a first observation (with horror) is that the data doesn’t appear to have been through any form of normalization. I reproduced it, using a first script, to verify.
1 248 740 (56.36%) of the 2 215 729 haven’t received a first dose in the data communicated.
966 989 (43.64%) - as previously underlined by Igor Chudov1 - of the subjects left are therefore the target of the current analysis.
These 43.64% of the subjects are representing 11 626 (31.15%) of the 37 315 deaths.
We are therefore talking, without any prior specification, of analyzing a sub-cohort representing less than 50% of the original population - itself significantly different in death rates of the general cohort - which further illustrates the terrible analytical quality of the dataset. We can quickly confirm the obvious, i.e. that this “dose 1 over age 50” cohort is sub-represented in deaths, using a chi-square test.
> chisq.test(matrix(c(11626,25689,955363,1223051),nrow=2))
Pearson's Chi-squared test with Yates' continuity correction
data: matrix(c(11626, 25689, 955363, 1223051), nrow = 2)
X-squared = 2405.1, df = 1, p-value < 2.2e-16This is explained, aside for the various bias mentioned, by the fact - detailed below - that the NZWB data doesn’t include most of the earliest doses administered in New-Zealand - while the most frail populations were prioritized.
2 331 074 of the 4 193 391 doses have been administered to the subjects having received a dose 1.
1 862 317 doses have been administered to the subjects not having received a dose 1.
These 43.64% of the subjects are therefore representing.. 55.58% of the doses, reflecting an obvious bias of increased tracking compared to the average of the cohort (2.41 dose per subjects with dose 1 against 1.49 dose per subjects without). Not only is this cohort more healthy than average, but it’s also more tracked.
Steve expected, it seems, an “on-the-fly” reply, without judging useful to precise any of this…
Breakdown by Age groups & focus on the sub-cohort of interest
The break-down of the ages at database export time, by groups, of this sub-cohort assumed to have received a dose 1, is the following.
Reproducing Steve’s chart results in the following.
Most of the deaths in this sub-cohort are occurring above 50 years old.
We can further zoom on the sub-cohort of interest, to explain the phenomenon observed by Steve Kirsch (the period framed in red), which consists in all the subjects whose age, at database export time, would have been over or equal 50 (a 313 744 subjects sub-cohort).
Normalizing by subjects available in each period of exposure (subjects who were still alive, have been recruited sufficiently early, and stayed alive long enough, to have been exposed to die within this time-frame), and considering a cut-off on September 30, 2023 to account for the reporting lag we already highlighted, results in the following.
Having smoothed the phenomenon already, it is now useful to put on a timeline the first doses administered & deaths in the sub-cohort, and in New Zealand.
Doses 1 & Deaths, in cohort & in New Zealand
First doses have been administered, in the sub-cohort - at a similar rate than the one observed in New-Zealand. We can observe that the NZWB sub-cohort (in green) hasn’t - as we mentioned above - received most of the earliest dose 1, administered in NZ (in red) to the populations prioritized, and most likely to die first (area circled in brown).
Deaths, as well, are fitting to New-Zealand’s global mortality.
The fact that the mortality grows, over the course of the year following the first doses administered, while the doses 1 are administered during a period of deaths decrease, is therefore a natural consequence of the New Zealand demographic trends which we already described in our previous articles. We can further illustrate it by looking at the overall mortality, in New-Zealand, compared to doses 1 administered.
Deaths & doses 1 administered in the 50+ sub-cohort can be represented as well as follows.
Simply put, the distribution of deaths to dose 1 within the year post said dose 1 is entirely related to the fact that the doses 1 have been administered prior a (predictable) spike in mortality. The dataset teaches us absolutely nothing which we didn’t already know, and could have modeled, from the New-Zealand published data.
We can further verify if their mortality fits the expected observed mortality in New-Zealand using a declination of the model we formerly used, focused on the New Zealand global data on these ages.
It demonstrates, once again, the under-reporting of deaths in the NZWB cohort, and overall behavior according to the expectations we can have, for this sub-cohort, from the background NZ mortality.
We are currently discussing with Steve Kirsch terms of an eventual debate. As I expressed him, I’ll be happy to proceed, on the premise that the one losing will go to retirement from public positions2.
The code used here has joined our NZ repository on GitHub.
When I kept people included under dose 1 even after subsequent doses, and when I calculated excess mortality based on the age composition of the cohort without adjusting for seasonal variation in mortality, then the total excess mortality up to September 2023 was about 109% for people who received the first dose in April 2021, 29% in May, -14% in June, -12% for July, -12% for August, 18% for September, and 51% for October: https://mongol-fi.github.io/moar.html#Effect_of_missing_doses_during_the_rollout_of_the_first_dose. Among the late vaccinees who received the first dose in September 2021 or later, there continued to be elevated excess mortality even in 2023.
The month with the most first doses given was August. Other doses also seem to have a similar "late vaccinee effect" where people who received the dose during the later part of the rollout peak later had higher excess mortality than people who received the dose during the earlier part of the rollout peak.
So there seems to be a distribution where first a small number of the earliest vaccinees have high mortality, second a large number of earlier vaccinees have low mortality, and third a large number of later vaccinees have high mortality. And the proportion of doses that are missing from the NZ data gradually gets lower over time, so the underrepresentation of the first group is counteracted by the overrepresentation of the third group.
Great job. I, too believe the bulk of mortality should be occurring within days.