Attached is a spreadsheet based on the data provided. After eliminating 4 instances with no information, there were 34 occurrences. Some takeaways:
- 18 unique TLDs represented in the 34 occurrences
- A median of 3 occurrences per TLD (that's to say that nearly half of them had 1 to 3 occurrences, the other half 4 to 7 occurrences)
- A median of 22 days between the delegation and the report, with some very late reporting (maximum was 568 days)
- 7 brand/exclusive-use TLDs, 1 GeoTLD/Governmental registry among the 18 TLDs
- 23 cases reported as service disruption, 9 as networking errors
- 21 cases affected company networks, 7 a single local computer, 2 application development environments, 1 web application
- In 24 cases the registry was not contacted, in 5 the registry was put in contact with the reporter, in 1 registry stopped controlled interruption, in 1 no action was taken
- All 5 known outcomes that were reported were that the network was updated, all others are unknown

</co-chair>

What was already known:
- No human-life threatening collisions
- Very low number of reported occurrences

What I expected and was confirmed:
- No IDN TLD collisions reported
- Brand TLDs having a share of the occurrences
- Little data on outcomes (people usually don't go back to tell their history after problem is fixed)
- Very few (1) instances where interruption was interrupted

What was unexpected to me:
- Very few (1) occurrences with GeoTLDs. It was forecasted that since many multi-site and global organisations built network hierarchies with city names and airport codes, those would have a higher incidence of collisions. Even some suspected cases like .cba, a brand, was supposed to generate collisions with geographic identifiers used in Chiba (JP), Córdoba (AR) and Cuiabá (BR). 
- Late reporting. Considering most people wouldn't wait almost 2 years to troubleshoot a problem, that indicates that some other conditions had to change for the collision to occur: actual registration of a colliding domain, network reconfiguration that allowed it to happen etc. 
- Reduced number of attempts to include registry into the incident response


Rubens





On Sep 26, 2017, at 8:30 PM, Steve Chan <steve.chan@icann.org> wrote:

Dear WT4 Members,
 
If you’ll recall, WT4 requested additional information for reports of name collisions. Below, please see the requested data elements.
 
  1. Date of report to ICANN
  2. Type of TLD where the collision occurred (Single-registrant, Brand, Geo, IDN, Open/Generic, Open/Niche)
  3. When and how reporting person detected the collision
  4. Affected system (Corporate network, Mobile Application, Web Application, Other-Specify)
  5. Registry response (If available)
  6. Outcome (to the best of ICANN's knowledge)
 
Attached, please find ICANN Org’s response (and available near the bottom of the Wiki page here: https://community.icann.org/x/Yz2AAw). 
 
Best,
Steve
 
 
 
 
 
Steven Chan
Policy Director, GNSO Support
 
ICANN
12025 Waterfront Drive, Suite 300
Los Angeles, CA 90094-2536
mobile: +1.310.339.4410
office tel: +1.310.301.5800
office fax: +1.310.823.8649
 
Find out more about the GNSO by taking our interactive courses and visiting the GNSO Newcomer pages.
 
Follow @GNSO on Twitter: https://twitter.com/ICANN_GNSO
Follow the GNSO on Facebook: https://www.facebook.com/icanngnso/
 
<Name collision report summary - 2017-09-26.xlsx>_______________________________________________
Gnso-newgtld-wg-wt4 mailing list
Gnso-newgtld-wg-wt4@icann.org
https://mm.icann.org/mailman/listinfo/gnso-newgtld-wg-wt4