刷题方法 错题本 检索练习 真题拆解
很多人把刷题当成体力活,题量上去了,分数却没什么变化。问题常常不在练得少,而在每一道题只走了一遍流程:看题、想一会儿、对答案、抄一句订正。下面这套做法把一百道题压缩成十道,每道题要过四个角度,做完的痕迹会留在纸上。省下来的时间用来整理和复述,而不是继续往前赶。
Most students measure practice by how many questions they finish. A better measure is how many questions were actually understood. Ten questions examined from four angles leave more behind than a hundred skimmed. The routine below keeps the workload small and the thinking slow. Each stage produces something visible on paper, so progress can be checked instead of felt. The time saved goes into review, not into the next worksheet.
题量错觉：重复步骤不等于重复思考
一套练习题的前半部分通常在教同一个动作,做到第十题左右流程就固定下来了,之后答对率上升只是熟练度,不是新的理解。很多学生以为刷完整本练习册就是努力,其实后面几十页只重复了最初的判断,边际收益低得可怜。判断一道题是否还值得做,看的是解题中有没有出现停顿、犹豫,或者需要回头翻公式的地方。如果从头到尾一气呵成,这道题已经完成了它的任务,换一道结构不同的才有新信息。
Within a single question type, returns fall away fast. The first ten items teach the pattern and the rest rehearse it. Accuracy improves, which feels like progress, while nothing new is being built. A question still has value only when some step makes you pause, guess, or reach back for a formula. When the whole solution runs without hesitation, that item has delivered what it had. Move to a different structure rather than repeating a comfortable one.
从考纲反推题型清单
AQA、Edexcel、OCR 的 specification 里有明确的考点列表,CAIE 的 syllabus 也按编号列出每个学习目标,这些列表就是刷题的边界。把最近三年的真题按考纲编号打上标签,很快能看出哪些编号每年都出现,哪些只露过一次脸。很多学生以为真题做完就结束了,其实标签表才是真正要建的东西,它决定下一轮从哪几道题开始。标签做完之后,整套题会自然缩到三四十道,剩下的题再刷也不会带来新的给分点。
The specification published by AQA, Edexcel or OCR lists every assessable point, and the CAIE syllabus numbers each learning objective. Tag past paper questions with those codes before deciding what to practise. Patterns show up within three years of papers: some codes carry several questions every session, others appear once. Most students stop when the paper has been marked, which leaves the map unbuilt. The map is the useful output, because it decides what to attempt next. Once the tags are in place, the working set shrinks to thirty or forty questions that actually carry marks.
合上书做检索练习
重读笔记时,眼睛扫过熟悉的内容会产生一种已经会了的错觉,这种错觉在考场上会立刻破功。合上书,把定义、公式、推导顺序默写出来,写不出的位置就是真正的漏洞,比再读十遍笔记都准确。常见的错误是把抄写当成复习,手指在动,脑子跟着原句滑行,提取的过程根本没有发生。每次提取之后立刻对答案,错的当场标记,当天再写一遍,这种方式的黏性远高于被动阅读。
Rereading notes creates fluency that does not survive an exam hall. The page looks familiar, so the brain files it as known. Closing the book and reproducing the definition, the formula, or the sequence of a derivation exposes what is actually missing. This is uncomfortable, which is exactly why it gets skipped. An extraction attempt followed by an immediate check is worth more than several passive passes over the same paragraph. Copying notes is the weakest version of review, because the hand moves while the mind reads along.
错题本按错因分三层记录
错题本最容易变成抄题本,一行题目加一行正确答案,复习时只能重新做一遍。把错因写在题目旁边,分成概念不清、步骤跳步、审题漏条件三档,复习时按类别集中处理。错得最多的是第二档,学生知道公式,却在中间某一步把符号或单位丢了,这类错误改起来最快。第三档要靠圈关键词的习惯,把否定词、单位、有效数字先划出来再动笔。
A mistake log degenerates into a copied question with the right answer beside it. That format forces a full re-solve every time it is opened. Label the cause instead: missing concept, skipped step, or misread condition. The middle category usually dominates, where the method is known and a sign or a unit disappears halfway. Fixing those is quick because the reasoning is already there. Misread conditions need a different habit, circling the command word, the unit, and the required precision before writing anything.
用 mark scheme 对齐给分点
考试局官网公布的 mark scheme 是按点给分的,答案对但过程缺一步,分数照样会扣。做完题后把自己的步骤和 mark scheme 逐条对照,数清楚每个方法分、每个答案分落在自己写的哪一行,这一步比再做五道新题有价值。很多学生以为分数只跟最终答案有关,翻开 mark scheme 才发现一个三分题要写三个独立判断。把常丢的点记在题号旁,下一轮同类型题先按这个框架写。
Mark schemes published by the exam boards award marks point by point, so a correct answer with a missing step still loses credit. Compare your written solution line by line with the scheme after every attempt. Count which of your lines earned each method mark and each accuracy mark. Five new questions teach less than one honest comparison of this kind. Three-mark items usually demand three separate judgements rather than one long paragraph. Note the marks you habitually drop, then write the next question of that type against that skeleton.
同一考点的三种问法变式
同一个考点在不同年份会换外壳,有的直接算,有的反过来给结果求条件,有的塞进一段实验或商业情境里。刷题时把这三类问法凑齐,比连做十道同向题更能检验理解,因为反向提问会暴露公式只会套不会用的地方。这一步最容易偷懒,学生做完正向后觉得已经掌握,遇到反解就卡在移项上。变式练完再回到真题,会发现题干再长也只是同一套动作。
One topic reappears in several disguises across years. Sometimes the value is asked directly, sometimes the outcome is given and the condition is wanted, sometimes the same mathematics is buried in a context. Collecting all three forms tests understanding far better than ten repeats of the first. Reverse questions expose a formula that was memorised but never rearranged. This is where students who felt confident suddenly stall on a single algebraic step. After the variations, a long past paper question looks like the same short routine.
间隔复习的日期安排
当天做对的题,隔一天再提取一次,隔一周再提取一次,隔一个月放进混合卷里再出现一次,这四次的成本比重做二十道新题低得多。间隔拉长的目的是让遗忘发生一点,再把它拉回来,提取越费力,留在脑子里的时间越长。常见的错误是把复习都堆在考前一周,那时只能重做,来不及重建。把日期直接写在错题本的页边,到期就做,不用每天决定复习什么。
A question solved correctly today is worth revisiting tomorrow, then a week later, then a month later inside a mixed set. Four spaced contacts cost less than twenty fresh questions and hold far better. The gap matters, since a little forgetting has to happen before recall does any work. Pushing all revision into the week before the exam removes that option, because there is no time left to rebuild. Write the next review date in the margin and stop deciding each evening what to revise.
限时训练与分段计时
平时不计时,考场上就会在会做的题上超时,最后两道大题空着交卷。限时训练先用整卷节奏跑一遍,再按板块分段计时,找出哪一个板块吃掉了多余的时间。很多学生以为慢是因为想得细,记录几次用时会发现,时间常常丢在读题和反复检查同一行上。分段计时之后,调整就有了具体对象,而不是笼统地说自己做得慢。
Untimed practice hides the real problem until the exam, where the last two long questions are left blank. Run a full paper on the clock first, then time each section separately to see which one swallows the minutes. The gap rarely sits where students expect. Time usually disappears in rereading the stem and rechecking the same line. Once the slow section is named, the fix becomes concrete: fewer passes, earlier commitment to a method.
基础题与难题的配比
把一轮练习分成三份,七成放在基础题和中等题,两成放变式,一成放压轴,这个配比能让分数稳定上升。基础题保证会的都对,难题拉开区分度,两者等量混在一起练,两头都不扎实。很多学生以为难题做不出来是能力问题,回看试卷会发现丢分大多发生在中等题的计算和单位上。基础段的正确率稳定在九成以上再往前推,进度反而更快。
Split a practice round into three parts: most of it on routine and mid-level questions, a smaller share on variations, the smallest on the hardest items. Routine work protects the marks already within reach, while hard items create separation. Mixing them evenly leaves both sides shaky. Students often read a weak score as a ceiling on ability, then find on review that the losses sit in mid-level arithmetic and units. Hold the routine section above ninety percent before pushing the difficulty up.
口头复述替代第十三遍重做
做完一道难题,合上答案,用自己的话把思路讲一遍,讲不顺的地方就是没真正想通的地方。讲给同学听效果更好,因为对方会追问为什么这一步成立,追问会逼出原本含糊的环节。常见的错误是把看懂答案当成会做,答案的逻辑是别人的,自己复述时才发现中间缺了一环。三分钟的口头复述通常比再做一遍更能暴露问题,而且不占用新题。
After a hard question, close the solution and explain the route out loud in your own words. Anything that comes out muddled was never really understood. Explaining to a classmate works better, because the follow-up question lands exactly on the vague step. Reading a worked solution feels like learning while the logic still belongs to someone else. Three minutes of speech exposes more than another silent attempt, and it costs no new questions.
正确率与用时双轴记录
只记正确率会漏掉一半信息,同一套题百分之八十的正确率,用六十分钟和用一百分钟代表两种完全不同的状态。把每次练习的正确率和用时并排写在一行,进步就不再是感觉,而是能看见的两条线。常见的错误是只盯着分数,忽略了时间在悄悄膨胀,等到考场才发现做不完。两轴都稳住之后再增加难度,顺序反了会同时失去速度和准确。
Accuracy alone hides half the picture. Eighty percent in sixty minutes and eighty percent in a hundred minutes are different results. Record both figures on the same line after every session, so progress becomes a pair of visible trends rather than a feeling. Watching the score while the clock quietly stretches is how students arrive at the exam unable to finish. Raise the difficulty only when both numbers hold.
每周删题与换题
错题本会越积越厚,不删就会在考前变成一座翻不完的山。每周复盘时,连续两次做对且能讲清思路的题划掉,剩下的按错因重新排队,新题只补进空出来的位置。很多学生以为错题本越全越安心,其实一本两百页的本子到考前只会被翻三页。删题的标准要提前写死,不然每次都会心软留下已经掌握的题。
A mistake log grows until it becomes a mountain nobody climbs before the exam. Once a week, cross out every question answered correctly twice and explained without hesitation. Re-sort what remains by cause, and only then add new items into the freed slots. A two-hundred-page log gets three pages of attention in the final month. Fix the deletion rule in advance, or every review keeps the comfortable questions and drops the useful ones.
考前三周把新题减到零
考前二十一天,新题的价值开始低于旧题,因为新题带来的是陌生感,而考场需要的是稳定输出。这段时间把同一批题反复提取,配合整卷限时,重点放在开局十分钟和收尾十五分钟的节奏上。很多学生以为考前要多见新题型才安心,实际上新题只会制造焦虑,已经掌握的题才是可靠的分数来源。
Three weeks out, new questions start to cost more than they return. Unfamiliarity creates noise, while the exam rewards steady output from material already owned. Recycle the same set through retrieval and full timed papers, with attention on the first ten minutes and the final fifteen. Chasing unseen question types at this stage mostly produces anxiety. Marks come from the questions you can already do under pressure.
牛津高阶英汉双解词典 第10版 商务印书馆¥168📱 去京东购买京东精选
需要一对一辅导？Need One-on-One Tutoring?
A-Level / GCSE / IB / KS2-KS3 全科辅导,按考纲拆分考点
16621398022 同微信
Visit aleveler.com
或截图后用「扫一扫 → 相册」选择本图