The error log that actually works
Nearly every student who has been told to keep an error log is keeping one that cannot work. The defect is structural, not effort.
A useful error log records the rule behind a miss and schedules its return.
An error log fails when it records questions instead of rules, and when it is a list instead of a queue. A list of forty questions you got wrong is an archive: it grows, it is never read, and it cannot tell you that six entries are the same rule. A queue records the rule, then schedules it back to you tomorrow, in three days, in a week, in three weeks.
Record the rule, not the question
The instinct is to copy the question, or screenshot it, and write what the right answer was. That produces an entry you cannot do anything with, because the question will never appear again and the answer was never the point. What generalizes is the rule, and the step at which you lost it.
So each entry is one line, written in your own words, naming the thing that will recur. Six fields, and none of them is the question text:
- The rule. “A comma alone cannot join two independent clauses.” Not “question 14 punctuation”.
- The category. Careless, timing, or not known. Three different problems with three unrelated repairs, and the sort has to be honest.
- The step that failed. Misread the stem, wrong setup, right setup and wrong arithmetic, solved for the wrong quantity, ran out of clock.
- The tell. What, next time, will let you recognize this is the same situation. This is the field that does the work and the one everybody leaves blank.
- Where it happened. Module one or module two, and whether under timing. A module-one loss is more expensive than a module-two loss.
- Next due date. Without this it is a list.
One more rule worth the discipline: log the questions you got right by elimination or by luck. They are not yours yet, they will not survive a form where the elimination fails, and they are the most reliable hidden cause of a score that stops moving.
Why most logs fail
- They are lists. A list only grows. Nothing in its structure ever brings an entry back, so review depends on willpower at a moment when there is no deadline.
- They store questions. A copied question is unreviewable. You already know the answer, so re-reading it rehearses recognition rather than the rule.
- They do not categorize. Without careless, timing and not known, every entry gets the same treatment, which is wrong for two of the three.
- They cannot show repeats. If entries are phrased as questions rather than rules, the same rule missed six times looks like six separate misfortunes.
- They are too long per entry. A log that takes five minutes an entry gets abandoned in week two. One line and a date survives a season.
The intervals
Review has to be spaced to stick, and the spacing matters more than the duration. Four passes, each doing a different job:
| Pass | When | What it is for | What a failure means |
|---|---|---|---|
| First | Next day | Confirming you can reproduce the rule without the explanation in front of you | You did not learn it; you recognized it. Re-teach, then restart the interval. |
| Second | Three days | Checking it survived a night and some interference | The rule is known but not yet retrievable. Keep it at three days. |
| Third | One week | Testing it on a different question of the same type, under timing | Known but slow. This is now a timing entry, not a knowledge one. |
| Fourth | Three weeks | Retiring it. Passing here means it leaves the queue. | Back to the next-day interval. It was never secure. |
The intervals are not precious to the day. Next day, three days, a week, three weeks is a shape that works and is easy to keep by hand. What matters is that the gaps widen and that a failure sends an entry backwards rather than forwards.
The smallest version that works
If you take nothing else from this: one page, one line per miss, the rule in your own words, a letter for the category, and a date. Sort the page by date each morning and do what is due. That beats any sophisticated system you stop using in February.
Every wrong answer explained in writing and sorted into a review schedule, across 4,495 questions and 20 adaptive practice tests. Both sections always included.
See what it costsBuild a free study planCommon questions
What should an SAT error log record?
The rule behind the miss, in your own words, rather than the question. Add the error category — careless, timing or not known — the step that failed, the tell that will let you recognize the situation next time, which module it happened in, and the date the entry is next due.
Why do error logs stop working?
Because they are lists. A list only grows and nothing in it ever brings an entry back, so review depends on willpower rather than structure. A queue schedules each rule to return, which is the difference between recording errors and repairing them.
What review intervals should I use?
Next day, three days, one week, three weeks. The gaps widen deliberately, and a failed pass sends the entry back to the start rather than on to the next interval. The exact days are not precious; the widening and the demotion are.
Should I write the question into my error log?
No. The question will not appear again and you already know its answer, so re-reading it rehearses recognition instead of the rule. Write the rule and the step that failed, then test them on a different question of the same type.
Should I log questions I answered correctly?
Log the ones you got right by elimination or by luck. They are not yours yet, they fail on forms where the elimination does not work, and they are the most common hidden cause of a score that has stopped moving.
Last reviewed . Exam format details are checked against College Board’s published specification each term.