Skip to main content
MarkovMarkov
Get the app

Help center

Self-service answers to the most-asked questions about Markov. If you can't find what you need, write to support@markov.in.

Understanding the different signal types

Last reviewed: 25 May 2026

The production stack

Six factor strategies run in parallel in production. Each one publishes its own signals into the feed, tagged with the originating strategy so you can see exposure mix at a glance.

Persistent quality

Looks for companies with consistent return-on-equity, low leverage and stable margins across rolling quarters. Long-only. Rebalances on a quarterly cadence — signals appear close to the rebalance windows.

Residual momentum

Strips out the broad-market (NIFTY) and sector betas from each name's total return; the stock-specific component left over is the "residual" momentum signal. Long-only. The idea: a stock outperforming its sector for stock-specific reasons is a cleaner signal than one that's just being lifted by sector tailwinds.

Sector-neutral momentum

Picks the strongest momentum name within each sector and weights them equally across sectors. The construction avoids the classic momentum failure mode of concentrating in whichever sector is currently hot.

Low-vol quality

Filters the quality universe down to names that are also trading with below-average realised volatility. Drawdown-aware: you give up some upside in exchange for a smoother ride.

Vol-targeted basket

Holds a basket and dynamically scales exposure to target a steady realised volatility band. Position size moves up when the basket is quiet and down when it's volatile.

Quality-momentum (QMJ)

The production workhorse. A composite of quality + momentum + the inverse of "junk" rank. Quarterly rebalance. This was the first strategy to clear our internal walk-forward and remains the highest- allocation strategy in the stack.

Reading the feed

Every card in the app carries a strategy tag. If you scroll the feed you'll see a mix of the six tags above. That mix is intentional — running multiple strategies in parallel diversifies the kind of signal exposure you take.

What's not in the production stack

Anything labelled "research-only" or "experimental" doesn't ship to the feed. We promote a new strategy to production only after it clears a multi-fold walk-forward gate set — typically 12+ out of 16 statistical gates.

Related in signals & methodology

Didn't solve your problem? Email support@markov.in and we'll get back inside one working day.