KraneShares CSI China Internet ETF
$28.09
-67.5%
Aug 12, 2026
C
MarginalGraded against KWEB's own history at this drawdown — not the market's.
What it means: KWEB's own history does not strongly favor buying at this level yet.
What to do: Consider watching and waiting for a deeper dip or a higher grade.
The fund will invest at least 80% of its net assets in instruments in its underlying index or in instruments that have economic characteristics similar to those in the underlying index. The index is designed to measure the equity market performance of investable publicly traded "China-based companies" whose primary business or businesses are in the Internet and Internet-related sectors, and are listed outside of Mainland China, as determined by the index provider. The fund is non-diversified.
Three plain-English paragraphs — current state, historical base rate, and interpretation — so you walk away with context, not a dashboard.
In plain terms: KWEB is 67.5% below its highest-ever price. Dipsern checks every past time it was this far down and grades how the next 90 days usually went.
KraneShares CSI China Internet ETF (KWEB) sits 67.5% off its peak and is currently in a historic crash, last printing at $28.09. Historic crashes (60%+ drawdowns) are rare events. For some assets they marked generational entries; for others they were warning signs of permanent capital loss.
Across all available history, KWEB has spent enough trading days in the 70%-to-65% drawdown band to build a large sample (328 observations). Dipsern segments these by drawdown depth and computes the rolling median forward 90-day return — that median is the primary signal you see in the full app, and it's substantially more robust than mean-based forecasts because it ignores fat-tail outliers in both directions.
For broad-index ETFs the historical base rate is exceptionally stable; for narrow thematic ETFs (e.g., ARKK, cannabis, leveraged), drawdowns are deeper and the sample size at each band is smaller. Crash-level drawdowns are tail events with few historical analogs. The signal should be read as "here's what happened the last few times" rather than a base-rate forecast.
Five tickers closest to KWEB's distance from all-time high. Same category, one tap away.
Dipsern segments the full drawdown range into 20 equal-width bands. The table shows the number of confirmed historical observations in each band for KWEB. Median forward returns, win rates, and prediction errors are available in the full analysis.
We segment 40+ years of KWEB's price history into 20 equal-width drawdown bands and compute forward median returns, win rates, and prediction error per band. No look-ahead bias. Pure NumPy. Open math, gated numbers.
Years of history
40+
Drawdown bands
20
Look-ahead bias
None
Data sources
4
See the Dipsern grade, median returns, win rates, prediction accuracy, 8 interactive charts, historical price paths, and a plain-English verdict for KWEB.
Dipsern is research and a second opinion — not a broker. You'd buy KWEB at your own broker.
Analyze KWEB FreeNo credit card required · 5 free analyses per month
Signal computed Aug 12, 2026 using default parameters (90-day return period, 20 segments, 0.95 decay). Data sourced from Yahoo Finance, Binance, CoinGecko, and Steam Market.