From multiple choice questions to “Shipitworthiness”
Under Review
SIRA is in the process of obtaining expert reviews of the model used and is also looking for guidance on whether the process IP should be protected before making the calculations public.
Both approaches help meet SIRA’s objective of promoting better information on innovation prospects. And both will help fulfil the aim to make disclosure and transparency an objective of the ratings activities. Your opinions and interests are appreciated:
Brief summary of how the probability rating is calculated
The assessment consists of two parts. Firstly, some data from questions on product development from the business plan, including financial ratios around growth. Secondly, a set of questions to obtain attributes which are up to 50 categorical answers on general product development challenges and approaches.
The latter are sufficiently generalised so that the model is not reliant on any specific dynamic system or simulation of R&D. And a “plug-in” section of the attributes is technology-specific. Nevertheless, using a type of component analysis, scores from the attributes can be converted into the probability of meeting the plan. By applying them to a statistically validated model and then using Bayesian Inference, the probability of meeting plan is consistently calculated. The Bayesian Inference means that the method can also benefit from improving accuracy as statistics accumulate.
Ratings and their probabilities
| Rating | Description | Probability |
|---|---|---|
| AAA | Prime | 0.98 |
| AA | High grade | 0.94 |
| A | Upper medium grade | 0.88 |
| BBB | Lower medium grade | 0.77 |
| BB | Investment grade speculative | 0.63 |
| B | Highly speculative | 0.37 |
| CCC | Substantial risks | 0.23 |
| CC | Extremely speculative | 0.12 |
| C | Low achievability | 0.06 |
| D | Not achievable | 0.02 |