---
title: "Notes on Packaging Alt Data, Signals, and Track Records"
description: "validityBase research and commentary on point-in-time data, predictive data, AI, and quantitative investing."
canonical_url: "https://www.vbase.com/blog/"
last_updated: "2026-10-15T10:39:24+00:00"
---

# Notes on Packaging Alt Data, Signals, and Track Records

## Articles

- [What Investors Want in a Data Provider’s Backtest](https://www.vbase.com/blog/what-investors-want-in-a-data-provider-backtest.md): The bar for a successful package is having the buyer feel “this is worth a serious look,” not “this is proven alpha.”
- [How Good Are LLMs at Sentiment Scoring?](https://www.vbase.com/blog/llm-sentiment-scoring-benchmark.md): We find that open-weight models, especially in an ensemble, approach frontier models’ performance. As with labeling, sentiment scoring is now accurate and cheap.
- [Если данные — нефть, то ИИ — сланцевая революция](https://www.vbase.com/ru/blog/dannye-neft-ii-slancevaya-revolyuciya.md): LLM снизили стоимость превращения неструктурированных данных в сигналы для инвесторов в 5–100 раз. Прежде невостребованные данные теперь могут стать ценным активом.
- [Почему кванты платят больше за point-in-time-данные](https://www.vbase.com/ru/blog/pochemu-kvanty-platyat-bolshe-za-pit-dannye.md): Квантовые инвесторы платят больше за проверяемые point-in-time-данные: только они позволяют достоверно восстановить прошлое и превратить бэктест в рабочее преимущество.
- [Почему проваливаются тесты альтернативных данных](https://www.vbase.com/ru/blog/pochemu-provalivayutsya-testy-alternativnyh-dannyh.md): Оценка альтернативных данных часто останавливается из-за их подготовки и предоставления, а не из-за отсутствия сигнала. Разбираем причины и решения.
- [Local LLMs can now handle ticker mapping for alternative data](https://www.vbase.com/blog/local-llm-ticker-mapping.md): Benchmarking open-weight LLMs for alt data pipelines
- [How good are LLMs at tickerization?](https://www.vbase.com/blog/llm-tickerization-benchmark.md): A reproducible benchmark comparing flagship LLMs and light LLMs in extracting asset tickers from unstructured text. Comparing performance of leading LLMs, including GPT, Claude, Gemini and Grok. We also discuss the pipeline…
- [If data is oil, AI is the shale revolution](https://www.vbase.com/blog/unstructured-data-shale-revolution.md): LLMs cut the cost of turning unstructured data into investor-ready signals by 5-100x. Data that was previously stranded may now be a valuable asset. This article explores what's changed and what options…
- [Your LLM’s alpha might be mere memorization](https://www.vbase.com/blog/llm-alpha-look-ahead-bias.md): A new benchmark, Look-Ahead-Bench, shows that LLM-based investment strategies can generate spectacular in-sample alpha that collapses once the test window moves past the model's training cutoff. For example, DeepSeek 3.2 went from…
- [Why alternative data trials fail](https://www.vbase.com/blog/why-alternative-data-trials-fail.md): Most alternative data evaluations fail before a fund even runs a backtest. This is not because the data lacks signal, but because it can't be evaluated with confidence. This article explains why…
- [Why quants pay more for point-in-time data](https://www.vbase.com/blog/why-quants-pay-more-for-point-in-time-data.md): Your data, made credible
- [Whose data is worth more?](https://www.vbase.com/blog/whose-data-is-worth-more.md): Your data, made credible
- [Stop using brokerage statements to show your track record](https://www.vbase.com/blog/stop-using-brokerage-statements.md): Your data, made credible
- [4 reasons people don’t trust your backtest](https://www.vbase.com/blog/4-reasons-people-dont-trust-your-backtest.md): Your data, made credible
- [4 reasons blockchains are excellent for data validation](https://www.vbase.com/blog/4-reasons-blockchains-are-excellent-for-data-validation.md): Your data, made credible
- [Who cares about data provenance?](https://www.vbase.com/blog/why-provenance-matters.md): Your data, made credible
- [Why great returns don’t attract investors](https://www.vbase.com/blog/why-great-returns-dont-attract-investors.md): Your data, made credible
- [Fundraising requires a track record](https://www.vbase.com/blog/fundraising-requires-a-track-record.md): Your data, made credible
- [Financial data must be made point-in-time](https://www.vbase.com/blog/financial-data-must-be-made-point-in-time.md): Your data, made credible
- [Why good data is hard to find](https://www.vbase.com/blog/why-good-data-is-hard-to-find.md): Your data, made credible
- [Beyond RFC-3161: the failures of legacy timestamping and a solution](https://www.vbase.com/blog/beyond-rfc-3161.md): Your data, made credible
- [3 reasons why GitHub timestamps shouldn’t be trusted](https://www.vbase.com/blog/github-timestamps-shouldnt-be-trusted.md): Your data, made credible
- [Is your pilot flying without a license?](https://www.vbase.com/blog/is-your-pilot-flying-without-a-license.md): Your data, made credible
- [Mitigating BorgBackup client compromise](https://www.vbase.com/blog/mitigating-borgbackup-client-compromise.md): Your data, made credible
- [Can invoice fraud be stopped?](https://www.vbase.com/blog/can-invoice-fraud-be-stopped.md): Your data, made credible