Evidence

Evidence for snyphr.com in the domain of insolvency prediction
On 9th November 2023 we imagined we were in April 2022
We made predictions on April 2022 data
There are two stages: calibration and backtesting
Calibration establishes settings achieving near-perfect prediction
Prediction is literal prediction

Calibration results

London firms had an insolvency rate of about 1%
Snyphr considered 10 samples of 100
1 true positive and 99 true negatives is a 2/100 chance, p=0.02
2 true positives and 98 true negatives is a 4/10000, p= 0.0004
Across 10 such samples Snyphr got 9 completely right
There was one false positive

Prediction back test

Considering firms reporting in Apr 2022
Snyphr ran ten times and predicted 10 firms to have an adverse outcome
7 of them were dissolved, liquidated, struck off, bought, etc.
3 were in distress with negative equity, etc.


These links are in alphabetical order

You might want to know about our business model.
business model

You might want to know about our coding discipline and open source.
coding and open source

You might want to know how to contact us.
contact

You could go to our home page.
company home

These are our products.
products

You might want to know about our staff.
staff

Here is snyphr's own page.
snyphr.com

You might want look at the tech and evidence outline.
technical evidence outline

You could take a look at our values.
values

Thanks again.