Sierra Chart's own engineers on Numbers Bars with Binance data

Ask Sierra Chart's own engineering account whether you can run a footprint chart on Binance data and you get a two-sentence answer with nothing hedged in it.

Numbers Bars cannot be used accurately with Binance data. And there is no trading support with Binance in Sierra Chart.

That is post #4 in the support board thread "Binance futures - DOM and FootPrint charts", posted by the account "Sierra_Chart Engineering" at 2025-10-11 20:21:30 and signed off in the usual block as Sierra Chart Support - Engineering Level. The user lion323 had asked whether Binance data is still reliable for order flow. Those two sentences are the whole reply.

Numbers Bars is Sierra's footprint chart, and it is a good one. Sierra Chart is a serious platform that is not in the habit of talking down its own studies, so when its engineering account tells a paying audience the study cannot be trusted on Binance, the reason matters. The interesting part is that they never gave one.

The map

Nobody got an explanation

The thread did not end at post #4. On 2025-10-12 22:16:26, a user posting as Cesium asked the obvious follow-up:

What part of the binance data is inaccurate for numbers bars?

The engineering account never came back. The only reply, from John - SC Support on 2025-10-13 18:36:20, was the phrase "Refer to the following:" and a link to Sierra's cryptocurrency data services documentation.

So anyone quoting this statement has to be careful. Sierra did not blame aggregated trades, missing tick data, or bid/ask classification, which is the usual suspect when a footprint goes wrong. They said none of those things. The reason has to be assembled from the page they pointed at, and that page turns out to be specific.

The answer is on the page they linked

From Sierra's cryptocurrency data services documentation, last modified 08 July 2025:

The volume data for the Binance crypto currency data feed is as integers only. It represents 10,000 times the reported fractional volumes from the original Binance data feed. The remaining digits to the right of the decimal place of a volume number after this multiplication, are then dropped.

And immediately after:

Therefore, the volume data should be considered inaccurate.

Read that arithmetic slowly. Multiply by 10,000, drop what is left of the decimal, and every trade is quantised to a grid of 0.0001 BTC, about $7.15 of notional at the price we recorded on the day of writing. Fills below the quantum contribute exactly zero volume. Larger fills are floored to the nearest step, so the error runs one way: recorded volume is never too high, only too low, and the shortfall piles up wherever size is being worked in small clips.

Numbers Bars is a volume-at-price study. It has one input that matters and this is it. A volume field the vendor itself calls inaccurate explains the engineering statement on its own, with no theory about tick reconstruction required.

The part that makes it fair

The same paragraph carries a sentence that most people quoting this would skip:

The reason for this handling is that Sierra Chart can only work with integer volume values.

That is Sierra describing their own constraint, in writing. The truncation is not something Binance does to the data. It is what happens when a storage format built around integer contract counts, the sane assumption for futures, ingests a market quoted to eight decimal places. Sierra picked a scaling factor of 10,000 and everything below it falls off the edge. The fault is shared, and to their credit Sierra says so on the same page they told the user to read.

The second failure mode

There is a second one on that page, nastier because it is silent:

With a very small Tick Size like .00000001, another problem that can occur is that there is an integer overflow causing inaccuracy with the Volume at Price data.

An overflow in the volume-at-price accumulator does not look like an error on screen. It looks like a number. Combine both mechanisms and the footprint may be short by every small fill and wrong at the extremes, with no visual tell either way.

Their view of the feed itself

Sierra is blunt about the source as well:

The Binance data feed has definitively proven to be a substandard, and unreliable data feed at its source. Outages and missing data are common. We do not provide any technical support for these issues at all. However, we do our best to restore outages, and patch missing data. It has been clearly identified that these issues are on the Binance side.

Note the shape of it. Outages and gaps are a supply problem; the integer truncation is a storage problem, and only the second comes with a documented mechanism.

Depth is not the reason

A tempting explanation, since Sierra also says market depth is not provided for Binance symbols:

Provided for BitMex symbols. Provided for Bitfinex symbols. Provided for Deribit symbols. Not provided for Binance symbols. Historical market depth data is not provided for any crypto currency symbol.

Tempting, and wrong. Sierra's own Numbers Bars documentation closes that off:

The Numbers Bars study has no dependency on Market Depth data. Numbers Bars only requires the best Bid and Ask price.

So the missing DOM is not what breaks the footprint. It does limit what you can build around one. Absorption is a claim about resting size holding while takers hit it, and spoofing is a claim about resting size vanishing before it can be hit. Neither question can be put to a feed with no book, and with no historical depth for any crypto symbol, neither can be studied after the fact.

The docs argue with the engineers

Here is the part that is genuinely confusing for a subscriber. The same cryptocurrency page also says:

The Numbers Bars, Volume by Price, and TPO Profile Chart studies can be used with any of the Cryptocurrency symbols provided by Sierra Chart. These studies use volume at price data.

Both statements were live on sierrachart.com on 2026-08-20. They reconcile if you read them as answers to different questions. The study runs and every cell fills. What Sierra distrusts is the number inside the cell, and a chart that renders confidently from a distrusted input is worse than one that refuses to draw.

FTX is still on the exchange list

One more detail from that page:

Set the Service to SC Data. You will have access to market data from the following cryptocurrency exchanges: Binance, Bitfinex, BitMEX, Deribit, FTX.

The page was last modified on 08 July 2025. FTX stopped operating as an exchange in November 2022. Bybit is absent from the list, as are OKX, Coinbase and Kraken. That is not a scandal, just a documentation page that has not kept up with the market it documents, and it shows where crypto sits in the priorities of a platform built for regulated futures.

What it costs to find out

Numbers Bars is not in the entry package. Package 3, Base Standard, runs 26 USD a month without the study. Package 5, Base Advanced, at 36 USD a month is the cheapest one that carries it, or 30.60 a month prepaid for three months, 27 for six, 23.40 for twelve. Package 11 is 46 and Package 12 is 56. On owning it outright:

Occasionally we get asked if we offer a one-time purchase option for Sierra Chart. The answer is no, and this is something that would never be offered.

There is a 21-day free trial with full functionality, after which the software drops to Free Mode, which does not include the crypto data services. Two conditions apply before you start that clock. The platform is Windows: "Sierra Chart is currently designed to run on Microsoft Windows operating systems", with Linux planned and no date given, and Mac users pointed at Wine. And the study needs the right storage format: the data records in the intraday chart data file need to be 1 Tick when using Numbers Bars, otherwise the displayed Numbers Bars data will be less accurate.

Then there is the structural limit. Asked on 2022-10-10, in support board thread 77391, whether two symbols could be combined into one Numbers Bars study, the same engineering account answered that this is not supported, and that you need two separate charts, each with its own study. For index futures that is barely a constraint, since one contract is the market. For crypto perpetuals it is the whole problem, because the market is split across venues that disagree with each other minute by minute.

What a merged tape looks like instead

We pulled our own aggregation layer at 14:38 UTC on 2026-08-20, mid-post, on a loud day rather than a quiet one. BTC perp last traded at 71,468.76, up 9.22% on 24 hours, on 68.86B USD of volume across the four venues, about 6.8 times the weekly median. Open interest stood at 16.50B USD: Binance 7.80B, Bybit 4.01B, Hyperliquid 2.53B, OKX 2.15B. Binance alone was 47.3% of it. Chart Binance and you have charted about half the positioning.

The 30-minute taker flow netted -82.63M USD, taker buy ratio 0.4545, while price slipped 0.45%. Split by venue: Binance -50.64M, Bybit -14.71M, OKX -9.38M, Hyperliquid -7.91M. Same direction everywhere, different sizes. The book disagreed outright: at that instant order book imbalance read +0.287 on Hyperliquid and +0.223 on Bybit, both bid-heavy, against -0.245 on OKX and -0.013 on Binance, both ask-heavy. Funding sat at the 85th percentile of the last 730 days. A one-venue footprint is not wrong in that state. It is answering a smaller question than the one you asked, and it cannot tell you which.

Our footprint chart merges Binance USDⓈ-M, Bybit linear, OKX and Hyperliquid into one cell grid across seven assets: BTC, ETH, SOL, BNB, XRP, DOGE and HYPE. The base cell is a 1-minute candle, aggregated client-side to 5m, 15m and 1h. Scrollback runs 30 days, the retention of the Parquet trade tape underneath, and live cells refresh about four times a second over WebSocket. Runs of 3 or more consecutive same-side imbalanced cells get a zone rectangle stroked around them, ratio selectable at 2:1, 3:1 or 4:1. Absorbed and pulled walls are classified every 30 seconds rather than eyeballed, and anchored VWAP takes up to 3 anchors that travel in the URL. How each of those is computed is written up in the footprint methodology; the chart's history, from the CBOT pit through MarketDelta and Sierra's own Numbers Bars, is in the footprint explainer.

What MarketTrace doesn't give you

Sierra Chart is a trading platform. We are not, and the gap runs wide in the directions a Sierra user cares about.

• No desktop application, no order routing, no broker connection. You cannot trade from our chart at all.
• No non-crypto markets: no CME futures, no equities, no FX, no options. Crypto perpetuals only, seven of them.
• No custom studies, no scripting layer, no backtester, no market replay.
• No DOM ladder, and no local tick storage you can query offline.
• Timeframes stop at both ends: 1-minute base cell at the bottom, 1h at the top. Nothing sub-minute, nothing multi-hour.
• Scrollback stops at 30 days, the retention of the trade tape underneath it.

If your work is futures order flow with execution attached, Sierra is the better tool and it is not close. If your work is crypto perpetual order flow, the decision gets made in the data layer, and that is the layer Sierra's own engineers told a user not to trust.

FAQ

Does Sierra Chart's footprint work with Binance data?

The study renders, but Sierra's own engineering account posted on 2025-10-11 that Numbers Bars cannot be used accurately with Binance data, and that there is no trading support with Binance in Sierra Chart. The post is two sentences long and carries no qualification.

Which Sierra Chart package includes Numbers Bars?

Package 5, Base Advanced, at 36 USD a month. Package 3, Base Standard, at 26 USD a month does not include it. Prepaying brings Package 5 down to 30.60 a month for three months, 27 for six, and 23.40 for twelve. There is no one-time purchase option and Sierra says there never will be.

Why did Sierra Chart say Binance data is inaccurate for Numbers Bars?

They never said in the thread. A user asked what part of the data is inaccurate and the engineering account did not reply; a support account posted a link to the cryptocurrency data services page instead. That page documents the mechanism: Sierra stores Binance volume as integers equal to 10,000 times the reported fractional volume, with the remaining decimals dropped, and states that the volume data should be considered inaccurate, because Sierra Chart can only work with integer volume values.

Can Sierra Chart merge several exchanges into one footprint?

No. Asked in 2022 about combining two symbols into a single Numbers Bars study, the engineering account replied that this is not supported and that you need two separate charts, each with its own Numbers Bars study.

Does Sierra Chart run on Mac or Linux?

Sierra Chart is currently designed to run on Microsoft Windows. Linux is listed as planned with no date attached, and Mac users are pointed at Wine.

Is there a free footprint chart for crypto perpetuals?

Yes. MarketTrace runs one at /perpetuals/footprint with no account and no install: Binance USDⓈ-M, Bybit linear, OKX and Hyperliquid merged into a single cell grid across seven assets, 1-minute base cells aggregated up to 1h, 30 days of scrollback, stacked imbalance zones and absorption markers.

Open one and check the cells

Every claim above about our own chart takes about fifteen seconds to check, and none of it needs a card or an account. Pick an asset, set the imbalance ratio, scroll back through the tape and see which venue was doing the buying.

Open the footprint chart →

Related

→ Footprint chart: from the CBOT pit to crypto perpetuals

→ Order flow trading: how to read the flow in crypto perpetual futures

→ Cumulative Volume Delta (CVD): the crypto perpetual futures trader's guide to order flow

→ Spoofing and fake walls: what the crypto order book is (and isn't) telling you

→ Where volume profile is actually free, and what each free tier hides

---

Sources

• Binance futures - DOM and FootPrint charts, support board thread 93271
• Cryptocurrency Data Services documentation, Sierra Chart
• Packages and pricing, Sierra Chart

---

The Sierra Chart threads and documentation pages quoted here were checked on 2026-08-20, and the MarketTrace figures were pulled from our own aggregation layer at 14:38 UTC the same day.