输入：建立任务模型
先确认题目要求的是立场、描述、比较还是总结，再标出必须覆盖的信息。模板只提供顺序，不替代当前题目的内容。
Identify whether the task requires a position, description, comparison, or summary. A template supplies order, not prompt-specific ideas.
一套可重复的训练闭环
A repeatable practice loop
每次练习都分为三个阶段。先理解任务与可用表达，再在时间限制内独立输出，最后根据真实错误安排下一轮。不要把看懂答案误认为已经掌握。
Each session has three stages: understand the task and useful language, produce independently under time pressure, then use real errors to plan the next attempt. Recognizing an answer is not the same as being able to retrieve it.
输出：在限制内完成
关闭不必要的提示，在真实时限内完成一轮。卡住时先保证答案完整，不要为追求一个漂亮句子而放弃后半段。
Remove unnecessary cues and complete one timed attempt. Preserve a complete response before polishing a single sentence.
复盘：选择一个变量
区分知识问题、注意力问题和时间问题。下一轮只调整一个高频变量，才能判断改变是否有效。
Separate knowledge, attention, and timing errors. Change one recurring variable so the next result remains interpretable.
五项能力应该怎样练
How to practise five skill areas
模板与打字 / Templates and typing
带参考输入用于熟悉句子结构，首字母提示用于主动提取，盲打用于检验是否能独立完成。速度只有在准确率稳定时才有意义；如果错误增加，先回到较慢的节奏并处理重复错词。
Guided typing introduces structure, initials require retrieval, and blind typing tests independence. Speed matters only when accuracy remains stable; if errors increase, slow down and review recurring weak words.
听力 / Listening
第一遍记录人物、主题、关键细节和结果，第二遍补数字、时态、介词与词尾。复述时优先保持信息关系正确，不必强行复制原句。
On the first pass, capture people, topic, details, and outcome. Use the second for numbers, tense, prepositions, and endings. In a summary, preserve relationships before exact wording.
阅读 / Reading
把错误标成词义、句法、篇章关系或时间压力。若同一词反复影响判断，将它加入错词复习；若单词都认识但句子仍读错，应重看从句边界与指代关系。
Label errors as vocabulary, syntax, discourse relationship, or time pressure. Repeated blocking words belong in review; familiar words with a wrong interpretation usually point to clause boundaries or reference.
写作 / Writing
先检查是否直接回应题目，再看理由是否有解释和例子。一个清楚展开的理由通常比三个泛泛观点更有价值。修改时保留原意，优先补足逻辑，而不是只替换“高级词”。
First check whether the response directly answers the prompt, then whether each reason has an explanation and example. One developed reason is usually stronger than three generic claims. Revise logic before swapping in impressive vocabulary.
口语 / Speaking
第一次录音先保证结构完整，第二次只改停顿、发音或细节中的一个。回放时不要只听口音；同时检查听者能否在前几句确定你的立场和回答方向。
Use the first recording to complete the structure and the second to change one issue—pausing, pronunciation, or detail. Do not judge accent alone; check whether a listener can identify your position early.
怎样理解站内指标
How to read the metrics
准确率用于定位拼写、遗漏、词尾与标点问题；速度用于比较准确率相近的练习；完成度用于检查回答是否覆盖任务要求；到期复习表示一次合适的主动回忆时机。
Accuracy highlights spelling, omissions, endings, and punctuation; speed compares attempts with similar accuracy; completion checks task coverage; and a due review marks a useful moment for active recall.
重要边界 / Important limitation
站内结果来自本地练习规则和个人记录，不是官方 DET 分数、考试认证或录取建议，也不保证固定幅度的提分。
In-site results come from local practice rules and personal records. They are not official DET scores, test certification, admissions advice, or a guarantee of a fixed score increase.
第一周：先建立基线
Week one: establish a baseline
每天 15–25 分钟，先完成五项各一次，再把第六天留给最低结果，第七天在连续计时下完成迷你模考。
Spend 15–25 minutes each day. Sample every skill once, reserve day six for the lowest result, and finish with a continuously timed mini mock.
DAY 1
模板打字
一轮带参考，一轮首字母提示。
DAY 2
听力
两道听写，复盘功能词和词尾。
DAY 3
阅读
为每个错误标记原因。
DAY 4
写作
先检查是否直接回应题目。
DAY 5
口语
同题录制两遍，第二遍只改一项。
DAY 6
重点复习
重做最低结果并处理到期错词。
DAY 7
迷你模考
连续计时完成四项并安排下周重点。
关于内容与独立性
Content and independence
TypeLoop 是面向 DET 学习者的独立练习项目。训练说明、界面反馈与学习路径由本站围绕实际练习流程整理，并根据产品测试持续修订。本站与 Duolingo 没有隶属、授权或合作关系。
TypeLoop is an independent practice project for DET learners. Its guidance, feedback, and study paths are organized around the product’s practice workflow and revised through ongoing testing. It is not affiliated with, authorized by, or endorsed by Duolingo.
词汇资料与第三方软件的来源和许可记录见项目中的第三方声明；用户生成的社区内容不代表本站观点。发现内容或功能问题时，可通过站内社区反馈。
Vocabulary resources and software licenses are documented in the project’s third-party notices. Community posts represent their authors, not the site. Product and content issues can be reported through the in-site community.
从一条真实记录开始
Start with one real attempt, then let evidence shape the next.