devwithjevJev-Powered Trading Backtest
WquGuru used OpenRouter’s Jev 1.13 to generate daily buy, sell, or hold decisions for a trading backtest, which returned +6.71% over 77 days.
WquGuru@wquguru𝕏
大开眼界，跑完交易回测算法，我见识到了 Jev 交易决策的效率和质量，况且只是简单地跑了几个算法，77 天收益率高达 +6.71%。 跑完之后剩一个问题：这一堆 JSON 怎么给人看？ 我把可视化设计稿丢给蚂蚁百灵刚开源的 Ling-3.0-flash-VL，让它照着写前端，最后输出视频。 我没给它设计稿标注，也没给它组件库。它自己看图，直接构建出动态仪表盘页面：净值曲线、三个策略的持仓条、当日决策流、右上角的加速倍数。然后它用 playwright 逐帧截图，390 帧压成 13 秒，浓缩了 95 天的新闻与行情。 说到具体的方法论：我把 NewsHub 攒了三个月的投资日报，一天一份丢给 OpenRouter 上的 Jev 1.13。我给它当天的日报和当前持仓，它能迅速输出每个标的买、卖还是不动，附带三个选项的概率和置信度。后面的程序拿着这些概率去调仓、按 Binance 真实小时级别
Jev received each day’s investment report and current holdings, then returned probabilities and confidence for each decision. A downstream program used the probabilities to rebalance. The author also used Ling-3.0-flash-VL to turn a design mockup into an animated dashboard video.
Also filed under Trading & markets
- Open Finance Decision Models
Eikos provides open decision models for finance with a Jev-compatible typed API and evaluation runners.
- AIStock Stock Analysis Application
AIStock is a stock-analysis application that uses Jev for structured market judgments.
- Intraday Trading Bot for Nifty 50
Warrenduffer is an intraday trading bot where Jev ranks Nifty 50 candidates, while deterministic code controls position sizing and risk.
- Crypto Market State and Risk Dashboard
A crypto decision dashboard uses Jev to classify market state and risk.