ac-predictor

Predict AtCoder's performance during the contest

您需要先安装一个扩展,例如 篡改猴Greasemonkey暴力猴,之后才能安装此脚本。

您需要先安装一个扩展,例如 篡改猴暴力猴,之后才能安装此脚本。

您需要先安装一个扩展,例如 篡改猴暴力猴,之后才能安装此脚本。

您需要先安装一个扩展,例如 篡改猴Userscripts ,之后才能安装此脚本。

您需要先安装一款用户脚本管理器扩展,例如 Tampermonkey,才能安装此脚本。

您需要先安装用户脚本管理器扩展后才能安装此脚本。

(我已经安装了用户脚本管理器,让我安装!)

您需要先安装一款用户样式管理器扩展,比如 Stylus,才能安装此样式。

您需要先安装一款用户样式管理器扩展,比如 Stylus,才能安装此样式。

您需要先安装一款用户样式管理器扩展,比如 Stylus,才能安装此样式。

您需要先安装一款用户样式管理器扩展后才能安装此样式。

您需要先安装一款用户样式管理器扩展后才能安装此样式。

您需要先安装一款用户样式管理器扩展后才能安装此样式。

(我已经安装了用户样式管理器,让我安装!)

作者
keymoon
日安装量
0
总安装量
17,405
评分
16 2 0
版本
2.0.12
创建于
2018-07-01
更新于
2025-11-09
大小
70.8 KB
许可证
MIT
适用于

Important Notice

For any issues, please report them as an issue on GitHub.

Regarding External Communication

This script communicates with external servers. The communication is strictly used for the following:

For detailed information about this data and why it must be fetched from external sources, please refer to the documentation. Since we only handle data retrieval, no account information or submission data is transmitted through this communication.

Details Technical details (as of December 2020) are available at The Technology Behind ac-predictor - Qiita. Below is a summary of sections relevant to external communication.

Why the Script Needs to Communicate with External Sites

For performance prediction, we require a value called APerf, which is calculated based on each participant's past performance.
If each user were to retrieve this value individually, it would result in approximately 10,000 requests per user. Considering load issues, we've implemented a system that aggregates these requests and computes APerf in batch, then distributes the results.

About APerf

APerf is a special rate used in performance calculations, derived from a weighted average of past performances. For detailed information, please refer to AtCoder's Rate Calculation Formula - Qiita. (This is the "internal rate" referred to in this article.)

About the Source Code

This script consists of approximately 1,000 lines of code. It was created using a tool to consolidate multiple files that originally existed as separate files. For those wishing to examine the script's source code, we recommend reading the pre-bundled source at this location. You can verify that running pnpm bundle at commit time for version updates produces identical script outputs.

Feature Description

Predictor

Adds a row to the rankings page displaying each participant's performance score and rating.

Estimator

Displays the value needed to achieve a specific rating or the rating you would have achieved with a particular performance score. The functionality can be toggled using the "Swap" button.