You're Not Weak at Evidence. You're Weak at 12 Specific Rules.
Vrenberg · June 20, 2026
Every bar prep tool tells you the same thing after a practice set. Civil Procedure: 68%. Evidence: 71%. Contracts: 74%. Congratulations, you now know that you have room to improve at Evidence. This is not useful.
The MBE does not test Evidence. It tests roughly 40 individual doctrinal rules within Evidence — hearsay exceptions, character evidence, impeachment, privileges, expert testimony, and so on. On any given practice set of 30 Evidence questions, you might see three hearsay exception questions, two character evidence questions, and one privileges question. If you got the two character evidence questions wrong and the rest right, your Evidence subscore is 93 percent. Fine.
But you have a character evidence problem. A subject-level dashboard will never tell you this. It hides the deficit inside an average.
The rule is the correct unit
Bar exam questions are not written at the subject level. Every MBE question is written to test either one specific rule or the intersection of two specific rules. NCBE question writers explicitly work from a rule outline when authoring items. The unit of the exam is the rule.
If the exam is written at the rule level, prep needs to happen at the rule level. Anything less granular is guessing.
Concretely: an evidence deficit that shows up as "71% on Evidence" is actually something like:
- Hearsay statement against interest: 40% (2 for 5)
- Present sense impression: 100% (3 for 3)
- Character for truthfulness: 25% (1 for 4)
- Prior consistent statements: 100% (2 for 2)
- Everything else: solid
The character truthfulness rule and the statement against interest rule are the entire deficit. If you spend a week reviewing "Evidence," you will spend most of that week on rules you already know. If you spend a week on those two rules, you fix the deficit.
Why every incumbent still tracks at the subject level
There are two reasons.
Their content was not tagged at the rule level. BarBri, Themis, and Kaplan have accumulated tens of thousands of questions over decades. Retrofitting rule-level tags across a legacy question library is expensive, tedious, and requires domain expertise to disambiguate ("is this a hearsay exception question or a hearsay definition question?"). The economic incentive to do it has been weak because the alternative — subject-level dashboards — has always been "good enough" for the market.
Subject-level UI is easier to design. A gauge showing 68% on Evidence is legible in one glance. A distribution showing your accuracy across 40 evidence rules is more information than most students want to look at. The design cost of making rule-level clarity comprehensible is real.
Neither reason is a good reason if you actually want to improve.
What rule-level tracking looks like in practice
If your prep tool tags each question with the specific rule it tests, then after a session you can see something like:
- Hearsay definition — 90%
- Excited utterance — 100%
- Present sense impression — 100%
- Statement against interest — 45% ← practice this
- Dying declaration — 80%
- Business records — 92%
- Character to prove conduct — 33% ← practice this
- Character for truthfulness — 50% ← practice this
- Impeachment by prior conviction — 88%
You have a clear list. Three rules to fix, not a whole subject to review.
The practice queue can then pull questions weighted toward those three rules. In a week, you can retake a diagnostic and see whether they moved.
The uncomfortable part
Building rule-level tracking requires knowing what the rules are. There is no NCBE-published canonical rule list for the MBE. Different prep companies have different taxonomies at different granularities. Some have 200 rules, some have 400, some just do subject-level tags and call it a day.
The taxonomy work is the hard, tedious part. It takes months of a domain expert going through the released question bank and annotating each question with the rule it actually tests. It is not glamorous. It does not scale by adding engineers.
We spent five months building a 391-rule MBE taxonomy from scratch. Each MBE rule is a page. Each rule tracks its own accuracy, its own last-attempt date, its own weakness score. When you get a question wrong, you know exactly which rule it was — and the next question the system queues is likely to be a related one, so you can address the misunderstanding while it is still fresh.
What to look for in any prep tool
If you are shopping for a tool other than ours, three questions to ask:
Does the tool tag each question with a specific doctrinal rule, not just a subject?
Can you see your accuracy per rule, not just per subject?
Does the practice queue re-prioritize based on rule-level weakness rather than subject-level averages?
If all three answers are yes, you have a tool worth using. If any answer is no, you are studying with a dashboard that will keep you from ever knowing what to actually fix.