# Golden cross and death cross on NSE stocks

> What the 50/200-day crosses actually mark, why they fire late by design, the whipsaw problem in sideways markets, and how to check their hit rate on NSE stocks.

Canonical: https://patternsradar.com/blog/golden-cross-death-cross-nse

Published 2026-08-15. Related scans: https://patternsradar.com/screener/golden-cross.md, https://patternsradar.com/screener/death-cross.md, https://patternsradar.com/screener/bounce-off-200.md, https://patternsradar.com/screener/stacked-averages.md

Few signals get more financial-media coverage than these two: the **golden cross** — a stock's 50-day moving average crossing above its 200-day — and its funeral-suited twin the **death cross**, the same lines crossing the other way. The names are theatrical. The mechanics are worth understanding precisely, because they explain both why the signals work and why they disappoint.

## What a cross actually records

A moving average is a summary of where price *has been* — the 50-day speaks for the last quarter, the 200-day for the last ten months. When the faster one overtakes the slower:

```
where sma(50) crosses above sma(200)
```

it records that **the recent regime has been better than the long-term regime for long enough to drag a quarter's average through a year's**. That takes months of net strength; a two-week rally cannot do it. Which yields the essential, honest property of the signal:

**It is late by design.** By the time a golden cross prints, the stock has typically already risen substantially off its low — the cross *confirms* a trend change, it does not catch one. Traders who treat it as an entry timing tool are disappointed on schedule. Its value is as a **regime marker**: which list a stock belongs to, not which day to buy it.

## Where each side earns its keep

**The golden cross as a filter.** The productive use is defining a universe: stocks that have printed a golden cross and hold above both averages are the pond where pullback-buying and breakout strategies actually work. The fully-developed version of that regime — price above the 20, above the 50, above the 200, in order — is the **[Stacked moving averages](https://patternsradar.com/screener/stacked-averages.md)** scan; the cross itself, on the day it fires, is **[Golden cross](https://patternsradar.com/screener/golden-cross.md)**.

**The death cross as a risk manager.** Its record as a *shorting* signal is mediocre — bear moves often end violently right around where the cross prints. Its record as a **"why are you still holding this?"** prompt is much better: a portfolio rule of reviewing anything that appears on the [death cross scan](https://patternsradar.com/screener/death-cross.md) costs little and occasionally saves a great deal.

## The whipsaw tax

The known failure mode: a sideways market walks the two averages into each other repeatedly, printing cross after cross that each reverse within weeks. Every crossover system pays this tax; the only questions are how often, and whether trend gains cover it. Two mitigations are standard — require price itself to confirm (`close > sma(200)` on the cross day), or use the related but earlier signal of price reclaiming the 200-day directly, which is the **[Bounce off the 200-day](https://patternsradar.com/screener/bounce-off-200.md)** scan.

## "Does it work?" is a measurable question

For NSE stocks, this year, the honest answer is not in any article — it is in the replay. Each scan page above carries a hit-rate panel that runs the scan across the past year of sessions and tracks what followed each signal. The pattern that typically emerges is exactly what the mechanics predict: crosses fire rarely, cluster at regime turns, look brilliant in trending years and pay whipsaw tax in flat ones.

Run [the golden cross](https://patternsradar.com/screener/golden-cross.md) replay and see which kind of year this one has been. Then, if the regime reading is favourable, do your actual stock-picking one level down — inside the [momentum family](https://patternsradar.com/scans/momentum.md), where the trend filters assume the regime and go hunting for entries.