This guide groups the skills the EQE Paper M3 catalog covers — drafting claims, replying to EPO communications, and preparing an opposition — around the decisions each one forces. Note that the official structure and administrative arrangements of the examination are set by the EPO; what follows teaches the underlying European Patent Convention skills, not the paper's official format. Every task below resolves into a visible choice: which feature defines scope, whether to argue or amend, which ground to plead and how to substantiate it. Each section isolates one recurring decision, shows a plausible mistake in a worked scenario, and states why the alternative holds up better. Practise by writing the reason next to the decision, because a division needs the reasoning, not just the result.
Selecting which features belong in the independent claim
The independent claim should carry only the features the brief presents as necessary to solve the stated technical problem. Preferred materials, dimensions, and optional extras belong in dependent claims that form the fallback structure.
Start from the technical problem the brief identifies, then test each candidate feature with one question: does the invention still solve that problem without it? Features that pass belong in the independent claim; features that fail are embodiment details. Keep two distinct concepts apart: clarity under Article 84 EPC asks whether the claim defines the invention precisely, while sufficiency under Article 83 EPC asks whether the skilled person can perform the invention across the whole claimed scope. A broad claim therefore needs a sufficiency check, not only a clarity check.
Draft in two-part form where the prior art position is known: state the features shared with the closest prior art in the preamble and the distinguishing features in the characterising portion. This discipline forces you to identify the contribution explicitly instead of burying it. Also weigh the client's commercial goal: a claim broad enough to catch a variant is worth more than a technically elegant narrow claim, so record why each feature was included. That written rationale becomes the argument you reuse when the claim is later attacked.
Building a claim set with workable fallback positions
Treat the claim set as a negotiation staged in advance: a main request at maximum defensible scope, then dependent claims that each reintroduce one feature, so any later amendment is a single, pre-supported step.
Order dependent claims so each adds a single feature and the ladder narrows logically. Watch two structural traps: dependent claims that contradict one another, and dependencies that merge alternatives into one claim. Where the description discloses several ways to implement a function, prefer generic wording that the description supports over listing alternatives inside a single claim, which invites clarity objections. Every dependent claim should also be traceable to a specific passage, because a fallback that cannot be shown in the filed text will not survive an added-matter objection.
Worked scenario: a brief describes a plant-watering device in which a soil-moisture sensor drives a valve, and discloses capacitive and resistive sensors plus a solar power option. A plausible mistake is drafting the main claim as 'a sensor that is capacitive or resistive, powered by a solar cell'. The better decision keeps the main claim to the generic sensor-and-valve combination, places each sensor type in its own dependent claim, and drops solar power to a further dependent. The first version merges alternatives and imports an optional feature into the scope-defining claim, shrinking protection for no validity gain.
Matching the reply to the type of communication received
Name the communication before writing. An examination communication with objections invites full argument or amendment; a communication indicating intention to grant narrows the options; a summons requires a reply within its time limit addressing every point.
The distinction that changes drafting most is between an examination communication raising objections and a communication indicating that the EPO intends to grant. Against objections you can argue, amend, or both, and a well-formed reply usually does both: argue the main request and file at least one auxiliary request. Against an intention to grant, the realistic options are narrower: approve the text, propose limited changes that stay within the filed disclosure, or request that particular claims be redrafted in a different form — so a reply written as a full inventive-step essay misreads the situation.
In opposition the same classification habit applies. The summons to oral proceedings usually carries the opposition division's preliminary opinion; a reply should respond point by point, file any amended request in clean and marked-up form, and submit evidence within the period set in the summons. The recurring drafting risk is the generic reply that repeats the notice of opposition without addressing the specific points raised — it gives the division nothing new to weigh and forfeits the chance to steer the discussion before the hearing.
| Communication | What it signals | Reply priorities | Typical drafting risk |
|---|---|---|---|
| Examination communication with objections | The division remains unconvinced on specific points | Argue the strongest points, file an auxiliary request, keep amendments supported by the filing | Answering one objection and leaving the others untouched |
| Communication indicating intention to grant | No remaining obstacle on the current text | Approve, or propose narrowly targeted amendments within the filed disclosure | Launching broad re-argument as if fresh objections existed |
| Summons to oral proceedings | The case will be decided at a hearing on a defined agenda | Reply within the summons period, address each preliminary point, file requests and evidence in time | Repeating earlier submissions without engaging the new opinion |
Answering inventive step objections with the problem-solution framework
Work the four steps in order: closest prior art, distinguishing features, their technical effect, and the resulting objective technical problem. A reply that skips a step forces the division to reconstruct your argument for you.
Novelty and inventive step differ in structure: novelty fails on a single prior-art disclosure that shows every claim feature, while inventive step can combine a document with common general knowledge or with another document. Within the problem-solution approach, dispute the choice of closest prior art only when you have a real reason, such as a different purpose or a missing structural link, because accepting it and winning at the problem reformulation is often the shorter path. Whatever objective problem you formulate, it must follow from the effect the distinguishing features actually produce.
Worked scenario: an application adds a vibration sensor that wakes a bicycle-light circuit from standby; the examiner objects over a document showing only a light sensor. A plausible mistake is the bare reply that 'D1 nowhere teaches a vibration sensor, so inventive step follows'. The better response identifies the effect — standby power saved — formulates the objective technical problem around it, then shows the prior art offers no pointer toward that problem. This matters because it gives the division a route to allow the claim and identifies which feature carries the effect if amendment is needed.
Keeping amendments inside the filed disclosure
An amendment is allowable when the skilled person, reading the application as a whole, directly and unambiguously derives it from the filing as filed. In opposition, amendments must additionally not extend the protection conferred.
Article 123(2) EPC targets added subject-matter. The boundary is not literal wording but disclosure: content that is implicit or necessarily present can be used, while a generalisation drawn from a single example usually cannot. The classic trap is merging two separately disclosed embodiments into one claim — each embodiment was disclosed, but their combination was not. Before committing to any amendment, run a three-way check: where does the feature appear in the filing, is it presented as essential or merely as an example, and does the amendment change the direction of the disclosure?
Article 123(3) EPC is the opposition-specific companion: amendments to a granted patent must not extend the protection it confers. Note how the tests differ — 123(2) compares against the application as filed, 123(3) against the granted claims. Micro-scenario: a patentee deletes a limiting feature from a granted independent claim to restore breadth. The plausible mistake is assuming deletion is always safe. The better decision checks whether the remaining features still delimit the invention; deleting a feature can broaden scope and therefore breach 123(3) even though nothing new was added.
Structuring an opposition so every ground is admissible and argued
Plead each ground under the correct head of Article 100 EPC — lack of novelty or inventive step, insufficiency, or added subject-matter — and pair it with the document and paragraph-level reasoning that substantiates it from the start.
Structure the notice of opposition so a reader can verify each attack independently: for novelty, identify the single document and map every claim feature to a passage; for inventive step, state which document is closest and what the combination is meant to contribute; for insufficiency, explain what the skilled person could not do; for added matter, quote both sides of the comparison. Grounds outside the Article 100 list do not by themselves form a basis for opposition, so an attack resting on one of them is space spent without a leg to stand on.
Worked scenario: an opponent believes a granted claim now recites a 'control unit' that the filed application described only as an optional component. A plausible mistake is asserting 'the feature was not disclosed as such' without citations. The better decision quotes the relevant passage of the application as filed alongside the granted claim and states precisely what was added and why the skilled person could not directly and unambiguously derive it. This matters because unsubstantiated grounds risk being treated as inadmissible, costing the opponent its attack before any discussion of the merits.
A practice routine with a self-check rubric and readiness gates
Alternate the three skill areas across study blocks: claim drafting from fresh briefs, replies to communications of different types, and opposition skeletons. Score each attempt against a short rubric and address the weakest criterion before moving on.
Exercise — essential-feature audit. Take any published application's description and list every disclosed feature in a table with three columns: the feature, the technical effect the brief attaches to it, and whether the stated problem remains solved without it. Then draft a main claim and three dependent claims from the table. Expected observations: at least two features marked as dispensable should end up in dependents rather than the main claim; alternatives disclosed side by side should appear in separate dependents; and if your main claim carries a preferred material or dimension, you have probably promoted an embodiment into scope.
Apply a rubric with learning milestones (they measure practice progress, not exam results): one, you can state the objective technical problem in a single sentence tied to an effect; two, every proposed amendment is traced to a specific filing passage; three, every opposition ground is mapped to its Article 100 head with evidence named; four, your reply matches the communication type. Sequence the work accordingly: short claim-drafting drills first, then argue-versus-amend decisions on objection communications, then opposition skeletons built from granted claim sets, finally mixed full tasks done against the clock and scored with the rubric. Adjust block length to the time you have.
- You can produce a defensible main claim from a new brief in one draft and state why each feature is essential.
- You can classify a communication on sight and list the reply options it actually leaves open.
- You can draft an opposition skeleton in which every ground names its document, passage, and Article 100 head.
- You can explain, for any amendment you make, whether Article 123(2) or 123(3) applies and how it is tested.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.