Market regimes: efficiency ratio, Fear & Greed and VIX

A market regime is the character of a stretch of prices: trending or ranging, calm or wild. Strategies are built for one character and hurt by another. A grid earns from a range and suffers in a one-way fall; a breakout strategy earns from a trend and bleeds in a range. Nobody knows the next regime in advance, but some readings describe the current one well, and a few are wired into A9's bots.

The efficiency ratio

The efficiency ratio (ER) asks one question: how much of the price's movement actually went somewhere? It divides the net move over a window by the total distance travelled along the way, counting every wiggle.

• A straight line scores 1.
• A path that wanders up and down and ends where it started scores close to 0.
• The same net move reached by zig-zagging scores in between: a path four times as long as its net move scores 0.25.

Unlike a moving-average cross, it needs no average to catch up. It measures the path directly, which is exactly what a grid fears.

The grid's direction gate

A new grid on A9 carries a gate built on the efficiency ratio. Every 15 minutes it measures the ratio over the last two hours, eight 15-minute blocks.

• If the ratio rises above 0.5 while the price is falling, the grid stops placing new buys. Its sells keep resting, so it can always exit.
• If the ratio rises above 0.5 while the price is climbing, it holds its sells, so a straight run is sold higher in one piece rather than line by line. Its buys keep resting.
• A hold is released only when the ratio drops back under 0.4. The gap stops the bot switching on and off every 15 minutes when the ratio hovers near 0.5.

On a sharp fall in XSOXL on 17 and 18 August 2026, the two-hour ratio first blocked buys with about 8% of the fall done. A 12/26 moving-average cross on 4-hour bars, the gate it replaced, first blocked with 92% of the fall done, and then kept blocking through the bounce, when a grid most wants to buy.

New grids also lock one side after three fills in a row on that side, until the price turns back by one line. Fills see a one-way move sooner than a two-hour window does. These are the form's defaults and can be changed or switched off on the bot.

The Fear & Greed Index

The Crypto Fear & Greed Index combines measures such as volatility, momentum and trading volume into one number from 0, extreme fear, to 100, extreme greed. A9 reads it from CoinMarketCap, with a fallback source, every 15 minutes.

A crypto bot can carry a Fear & Greed floor. While the index is below it, the bot opens no new buys; exits are never blocked, because a gate that traps a position is worse than none. A floor of 20 corresponds to extreme fear. On a martingale the floor stops new cycles from starting but never blocks the extra buys of a cycle already running, since refusing the buy that lowers the average makes a bad cycle worse. If the index cannot be read, the gate stays open rather than stranding the bot. Stock bots do not use it at all: crypto sentiment says nothing about a share.

VIX, and why the gate was removed

The VIX is the volatility the options market expects for the S&P 500 over the next month. A9 shows it as calm under 15, normal under 20, elevated under 30, and high fear from 30. Over the ten years to August 2026 its median was about 17.

Until August 2026, A9's stock bots refused new buys while the VIX was above 35. The gate was removed on 18 August 2026, for reasons worth knowing if you watch the VIX yourself:

• It rarely fired. Over ten years the VIX closed above 35 on 2.43% of trading days, and not once between April 2025 and the removal.
• When it did fire, it could not tell a scare from a crash. The typical episode lasted two trading days; the longest lasted 36.
• One number for every stock bot cannot serve opposite strategies: a grid and a breakout want different things from high volatility.

The grid's direction gate replaced it, because it reads the instrument's own price, per bot. The VIX stays on A9 as a reading.

The technicals rating

Each stock's market page shows a rating from strong sell to strong buy. It tallies daily readings: the price against twelve moving averages (simple and exponential, from 5 to 200 days), RSI(14) and MACD. Each votes up, down or neutral, and the balance of votes sets the rating. A9 maps it to the strategy tags: buy leans to trend strategies, neutral to range strategies, sell to accumulation. It is a description of the recent past and has not been tested as a signal; it is a place to start choosing, not a forecast.

Using these readings

• Choose a strategy whose weak regime you can tolerate, rather than one that needs you to predict the next regime.
• Let the bot's own gates do the regime switching they were built for, and read their state on the bot page.
• Pausing a bot is a tool for when conditions have left what the strategy was designed for, such as a grid's range broken by a vertical move. It is a poor tool for timing the market: a resume after a missed move is usually a worse entry than the bot would have chosen.