#Kroniq Research

#The Intelligence of Market Timing

Real-time market regime detection infrastructure for quantitative finance. Building the API that tells every quant fund what the market is actually doing.

#Week 1 — Python for Finance (March 26–30, 2026)

#What has been built so far

File	Description
kroniq_week1_spy.ipynb	SPY 5-year price chart — first data pull
kroniq_week1_5asset_data.ipynb	5-asset data pipeline (SPY, QQQ, VIX, TLT, GLD)
kroniq_5asset_prices.csv	10 years daily closing prices — 2,765 rows
kroniq_5asset_returns.csv	Daily percentage returns — all 5 assets
kroniq_5asset_volatility.png	Rolling 20-day volatility chart with regime events

#Research thesis

Quant algorithms are blind to market regime shifts. When the regime changes, strategies built for one regime destroy capital in another. Kroniq detects regime shifts in real time using cross-asset Hidden Markov Models.

#Stack

#Week 2 — Quant Risk Analysis (Mar 30 – Apr 3) | File | Key Result | |------|-----------| | kroniq_week2_montecarlo.ipynb | Sharpe 0.5308 · Drawdown -33.72% | | kroniq_week2_var_cvar.ipynb | VaR -1.93% · CVaR -3.20% · 11.58% gap | | kroniq_week2_regime_routing.ipynb | 7×6 routing table — Engine 2 |

#Founder

Sumanth Polavarapu — Data Engineer @ Capital One. Building Kroniq: the AWS of quantitative finance.

#Week 3 Results (April 6-12, 2026)

#Model selection — BIC sweep K=3..9

K	BIC	Min state days	Decision
3	24,724.7	—	Rejected
4	23,852.4	—	Rejected
5	22,867.2	233 days	★ Production model
6	23,035.0	24 days	Rejected (unstable)
7	23,048.8	9 days	Rejected (unstable)
8	22,521.7	13 days	Rejected (unstable)
9	22,720.8	—	Rejected

#v4 Production model (K=5, 9 features)

Static OOS Sharpe: 1.107 | MaxDD: -8.37% | CAGR: 7.94%
Walk-forward OOS Sharpe: 0.881 | MaxDD: -17.46% | CAGR: 9.97%
Buy-and-hold Sharpe: 0.859
141-day COVID early warning (credit spreads)
All 3 sanity checks pass

#Transition matrix persistence

Low-Vol: 95.5% | Bull: 92.7% | Neutral: 92.5% | Macro: 81.7% | Crisis: 73.0%