Operating principles
Show the math
Every composite is a weighted average of named, observable inputs. No black boxes, no proprietary mystery weights.
Calibrated, not cosmetic
Weights vary by sector and fundamental classification. A REIT and a growth software company are not scored on the same yardstick.
Conservative under uncertainty
When data is missing or fails a check, the engines do not invent it. A valuation model without its core inputs gets no weight, a few secondary inputs such as the discount rate fall back to fixed default assumptions, a Score pillar without data is set to neutral or left out with lower confidence, and a fair value that fails its checks is not published.
No buy or sell language
A score is a snapshot of conditions, not advice. We display scores, classifications and fair-value gaps; the decision belongs to the investor.
MrChart Score
A score from 1 to 10 for each stock, built from four pillars: technical, fundamental, value and risk. The pillar weights follow the stock's classification, so a high score on a growth stock is earned differently from a high score on a value stock.
The four pillars
- Technical: trend against the moving averages, momentum (RSI and MACD), volume, and where the price sits within its Fibonacci retracement levels.
- Fundamental: valuation (P/E against the sector), growth, profitability and balance-sheet strength.
- Value: how far today's price sits from the MrChart Value engine's value, before it is averaged with the analysts' consensus (see MrChart Value below).
- Risk: volatility (beta and average true range) and how close the next catalysts are, such as earnings, Fed meetings and major economic releases. A higher Risk pillar means lower risk.
How pillars combine
The Score is a weighted average of the four pillars, kept between 1 and 10 and shown to one decimal place. The weights follow the stock's classification: growth stocks lean on the technical pillar, value stocks on fundamentals, turnarounds on risk, and blue chips balance fundamentals with technicals.
Fallback behavior
When the MrChart Value engine publishes no value for a stock, the Value pillar is set to the neutral midpoint and confidence is lowered. When technical, fundamental or risk data is missing, that pillar is left out, the remaining weights are rebalanced and confidence drops. With no usable pillar, no Score is shown.
MrChart Value
Eight independent valuation models, weighted by company type and life stage and checked against peers, produce the engine's value. The published fair value is the average of that value and the analysts' consensus target when one is available, compared with today's price.
The eight models
Discounted cash flow and total payout
Discounted cash flow projects the business forward, adds a terminal value and discounts both at the company's cost of capital. Total payout values the dividends and net buybacks the company returns to shareholders.
Forward P/E and growth-adjusted P/E
Forward P/E applies the company's own historical P/E to its forward earnings. The growth model sets a fair multiple from expected earnings growth and the peers' PEG ratio.
Peer multiples and price to sales
Peer multiples value the company on the median multiple of comparable companies, using the multiple that suits its type, such as P/E, EV/EBITDA or price to sales. Price to sales applies the company's own historical sales multiple to forward revenue, for companies priced on future growth rather than current earnings.
Asset-based and P/FFO
Asset-based value applies the peers' price-to-book multiple to book value per share, for balance-sheet businesses such as banks, insurers and real estate. P/FFO values REITs on their funds from operations.
Weighting and checks
Each company is classified by type, such as bank, REIT, utility, cyclical or profitable growth, and by life stage: emerging, expansion, mature or decline. The type sets each model's weight and the life stage shifts weight between the growth and cash-flow models. Earnings and peer multiples are adjusted for business quality (margins, return on capital, stability and leverage), peer multiples feed several of the models, and a model that does not fit the business or lacks its core data gets no weight. The analysts' consensus target also enters the engine as an outside reference: its weight grows when the models disagree with each other or with the consensus, and shrinks when few analysts cover the stock, their targets are old or they are far apart.
MrChart Flow
Dealer positioning and whale levels made readable, using the same mechanics institutional desks rely on. Where dealers are positioned shows where price is likely to meet support and resistance.
The Greeks we surface
- Gamma Exposure (GEX): net dealer gamma per strike. Positive GEX dampens volatility, negative amplifies it.
- Delta Exposure (DEX): directional dealer hedging pressure.
- Vanna: sensitivity of delta to implied volatility shifts. Skews dealer hedging when vol moves.
- Charm: delta decay over time. Drives end-of-day and end-of-week hedging flows.
From the option chain to levels
Strike-level open interest and options activity are turned into the levels that matter: gamma levels, the call wall and put wall, the gamma flip and whale levels, alongside whale trades and Dark Pool activity, for every stock and for the whole market, refreshed through each session.
MrChart AI Agents
A team of 11 specialist AI agents working on top of the same engines. Ask a question and the agents gather the evidence, debate the bull and bear cases, review risk and check every figure before answering.
The agents do not invent numbers. They read the same MrChart Score, MrChart Value and MrChart Flow data as the rest of the platform. Five agents gather and analyze the evidence (market data, technical, fundamentals, news and sentiment), a bull researcher and a bear researcher debate it, a research manager weighs the two cases, a risk committee and a risk review examine the downside, and an evidence check confirms that every figure in the answer is backed by platform data. The answer never ends with a buy or sell recommendation.
Language is enforced
A language sentinel locks each response to the question's language. We never sprinkle the other language into the answer for flair.
What we do not claim
- We do not predict short-term price moves. The Score is a conditions snapshot, not a forecast.
- We do not give buy or sell instructions. Investment decisions belong to the investor.
- We do not extrapolate confident scores from thin data. Missing inputs lower the Score's confidence or leave a figure unpublished.
- We do not promise the engines are right. They are calibrated, conservative, and observable, but markets remain uncertain.