Trarian Invalidity Score™ · Sample report
Patent invalidity assessment
For underwriters, brokers, insurers, lenders, & portfolio owners
Run ID TR‑684 v1.3 · 24h · σ = 0 Issued 21 May 2026
Prepared forSample · illustrative
Model versionTrarian v1.3 · GBM + LR
Training cohortn = 1,296 · PTAB + Fed. Cir. + ITC
Held‑out AUC0.791
Patent
Quality of service in a voice over IP telephone system
Number
US 7,068,684 B1 — issued 27 June 2006
Priority
14 February 2001 · CIP of US 09/783,932
Assignee
Estech Systems, Inc. (Plano, TX)
Examiner
Wing F. Chan · Art Unit 2664 · 14.2 hrs on application
Family
3 members · 1 US continuation, 2 foreign · forward cites n = 41
Status
Granted · maintenance fees current · subject of PTAB IPR2024‑00417

Light examiner search. Crowded prior‑art neighborhood. The bet is on invalidation.

TR‑684 lands in band 7 of the Trarian invalidity scale (1 = safest, 10 = riskiest). Patents historically scoring in this band have been invalidated at a rate of 65% when challenged on the merits — a band an underwriter can size positions against.

The driver story is coherent: the examiner added essentially no prior art of their own during prosecution, the citations that are on the file are themselves heavily cited (a busy technical neighborhood), and the patent sits in a contested forward‑citation pattern. Doctrine load on §103 is moderate but not damning.

0.791
model AUC · held‑out
65%
band 7 historical rate
σ = 0
run‑to‑run drift
Where TR‑684 sits in the distribution
historical invalidity by band · n = 1,296
100%
75%
50%
25%
7.7%
26.2%
25.6%
33.8%
50.4%
66.2%
65.1%
72.3%
79.1%
87.7%
1
2
3
4
5
6
7
8
9
10
Top drivers · signed contribution to invalidity
5 of 312 features shown
01
Examiner added essentially no prior art of their own during prosecution. Source · prosecution · file history
+0.184
02
References on the file are themselves heavily cited (busy, well‑mapped technical field). Source · forward‑citation graph
+0.151
03
Small patent family · fast‑moving forward‑citation pattern suggests contested space. Source · family · citation velocity
+0.118
04
Independent claim breadth: moderate. §102 anticipation risk: low. §103 obviousness load: moderate. Source · LLM claim analysis
−0.071
05
§101 eligibility risk: low. Specification‑to‑claim support: strong. Written description: clean. Source · LLM specification read
−0.098
Independent claim 1 · mapping to closest reference
US 6,452,924 B1 (Golden, 2002) · sept 2002 issue
Elt.Claim textDisclosure
1[a] "a method of providing quality of service in a voice over IP telephone system" Golden col. 2, ll. 14–28verbatim
1[b] "detecting a packet loss event on the network" Golden col. 4, ll. 31–52verbatim
1[c] "adjusting a buffer depth at a receiving endpoint in response to the detected loss" Golden col. 5, ll. 8–22 · cf. Fig. 3Bteaches
1[d] "signalling the adjustment to a corresponding transmitting endpoint over the same channel" Golden col. 6, ll. 41–58 · Yu §3.2obvious to combine
1[e] "wherein the adjustment maintains call latency below a configurable threshold" Golden col. 7, ll. 12–24verbatim
Closest prior art · top 6 of 142 surfaced
100+ authorities · patent + non‑patent · ranked by relevance
ReferenceDateDisclosureTheory
US 6,452,924 B1 Top hit 17 Sep 2002 Golden — jitter‑buffer adaptation in VoIP endpoints. Discloses 1[a]–[c] and [e] verbatim. §102 anticipation
US 6,389,032 B1 14 May 2002 Cohen — endpoint‑to‑endpoint signalling protocol for QoS feedback. §103 (w/ Golden)
IETF RFC 3550 §6.4.1 Jul 2003 RTP control protocol — receiver reports of loss rate, jitter, signalling. §103 (w/ Golden)
EP 1 209 879 A2 29 May 2002 Nortel — adaptive packetization rate to maintain latency. §103 combination
Yu et al., ICC 1999 Jun 1999 "Adaptive Playout for Internet Telephony" — §3.2 teaches mutual signalling. §103 motivation
Cisco AVVID 4.0 Manual Jan 2001 Commercial deployment manual — POSITA had ready access to the technique. POSITA context
Outcome · back‑test

The bet would have paid off.

TR‑684 is a real, settled case — included here as a worked example. The PTAB cancelled the challenged claims; the Federal Circuit affirmed. The score was made blind to the outcome.

A single case never validates a model. But it is the kind of outcome the score is built to flag in advance — and the kind of bet an underwriter can size to the band.

— In re Estech Systems IP, LLC, No. 24‑1935 (Fed. Cir. 23 Dec 2025) · affirming PTAB IPR2024‑00417

Caveats & reading notes

The Trarian score is a ranking tool calibrated to historical outcomes. It is not a guarantee of future results. In any small sample you will see patents in the strongest bucket that turn out invalid and patents in the weakest bucket that turn out valid.

We classify patents into more or less likely to be invalidated — not into "valid" and "invalid." A 1 is not a verdict; a 10 is not a free pass. The point is consistency: a 6 means the same thing every time, every patent, every analyst.

This report is a sample. Names and selections are illustrative. Trarian is a research firm, not a law firm. Nothing in this report constitutes legal advice.

Trarian. · Chicago, IL · hello@trarian.co
TR‑684 · v1.3 · pp. 1/1