FAQ
Questions, answered
From getting a backtest right to buying a licence. Each question has its own page, with the short answer first and the detail after it.
Getting systematic research right
Why don't my backtest results match live trading?Usually because the backtest assumed cheaper costs and easier fills than the market gives, used information that was not available at the time, or tested a slightly different strategy from the one that went live. Each cause can be found and closed before real money is at stake.My strategy was profitable in the backtest but loses money live. How do I find out why?Rerun the backtest over the exact dates you traded live, then match each live trade to its backtest trade. Missing or extra trades point to the code or the data. Same trades at worse prices point to spread, slippage and costs. If the backtest also lost money over those dates, the market changed or the strategy was never robust. If you have only a few trades, the loss may still be normal variation, so count them before you change anything.Why is my Python backtest so slow?Usually because a Python loop does expensive work on every bar, and that cost is multiplied by every bar and every run. Reading a value through pandas inside the loop is often the largest part. In one test, reading a single column row by row took about 12 seconds per 500,000 bars with iterrows and about 0.06 seconds with a NumPy array. Profile first, move data out of pandas before the loop, precompute what you can, run sweeps in separate processes, and move the loop to compiled code only if that is still not enough.How do I test a trading system before I risk real money or buy a prop firm challenge?Test it in steps, each of which catches a different way a system fails: a backtest with your real costs, a trade-by-trade inspection, a run on data you did not tune on, a Monte Carlo of the trade order, and paper trading at your broker with the same orders. Reamer Research, a Python and C++ backtesting and live trading engine, runs every step on your own machine, with byte-identical results, and its relay sends the orders you tested to a paper account and then a live one.How do I know if my backtest is overfit?You cannot prove a backtest is not overfit, but you can test for the signs. An overfit result holds at one parameter set and collapses at its neighbours, fails on data it was not tuned on, rests on a handful of trades, or disappears once realistic costs are applied. Check each one, on a tool that gives the same answer every run.Why do I get different backtest results each time I run it?Because something in the run is not fixed. Usually it is unseeded randomness, data that changed between runs, an order of operations that is not guaranteed, or a library or setting that drifted. Run the same backtest twice, compare the trade logs, and find the first trade that differs. A seeded engine gives byte-identical output, which leaves your own code and data as the only things to check.How realistic do slippage and spread need to be in a backtest?Realistic enough that the result survives them. Set spread, slippage and commission per instrument from what your broker actually charges, let spread and slippage widen when the market moves fast, and then rerun at one and a half and two times those costs. A strategy whose edge disappears under slightly higher costs was never a reliable edge.How do I manage many research scripts without results drifting apart?Separate what changes between studies from what must not. Each study should contain only its strategy logic, while fills, costs, data handling and metrics live in one shared, versioned engine that every study calls. Record the engine version, configuration, seed and data with every result, so any two results can be compared like for like.Can one laptop run serious quant research and live trading?For mid-frequency strategies on bars, yes. Research on a 2017 four-core laptop can process hundreds of thousands of bars per second, and a live order pipeline on the same laptop can keep latency in the tens of microseconds with one strategy per core. The limits are cores, uptime and the network, not raw speed. Keep live strategies on their own cores and plan for the machine to restart.What does professional-grade trading infrastructure actually include?Less hardware than people expect and more discipline. Research that gives the same result every run, fill and cost rules written down, risk checks kept separate from strategy code, one ordered path for every order, a record of every decision, live monitoring, safe restarts, and performance claims backed by published data. None of it requires a large team or a data centre.
Choosing the right kind of tool
What makes a backtest deterministic, and why does it matter?A backtest is deterministic when the same strategy, data, settings and seed give the same output, byte for byte, on every run. That requires fixed randomness, a guaranteed order of events, and nothing that depends on the clock, the machine or a previous run. It matters because every comparison depends on it: when the tool adds no variation of its own, any difference between two results comes from something you changed, and you can check what.What is a quant research engine, and how is it different from a backtesting script?A research engine is one shared program that does everything around a strategy (loading and aligning data, simulating fills and costs, accounting, metrics, logs and reports) so that each study contains only the strategy itself. A backtesting script does all of that on its own, once per study. The engine is what lets you test, diagnose, sweep, report and move to live with the same rules every time. A script is quicker to start and is the right tool for a single rough idea.Should I build my own backtester or buy one?Build the parts where your edge lives and buy the parts it does not depend on. For most bar-based strategies the edge is in the signals, the data and the sizing, not in how fills, costs and accounting are simulated, so buying or adopting an engine and building the strategy is usually the better use of time. Build the engine yourself when the engine is the edge, when nothing available models what you trade, or when you need to own every line of it.What software do independent quants use to go from research to live trading?A small stack of five layers: market data, a research engine, an execution pipeline that holds order state and runs pre-trade risk checks, a broker connection, and monitoring with a durable record. Few people buy all five from one place. The usual pattern is to buy the data and the brokerage, buy or adopt the engines, and write the strategy, the risk rules and the glue between the layers yourself.What is an order management engine, and does a small trading operation need one?An order management engine sits between your strategies and your broker. It tracks the state of every order, puts orders from every strategy into one sequence, runs a pre-trade check on each one, sends it to the broker, and records every decision and fill. Every trading system already does these jobs somewhere, often scattered through strategy code. A small operation needs a dedicated engine once more than one strategy trades an account, once real money is at stake, or once the system runs without someone watching it.Where should pre-trade risk checks sit in a trading system?Between the strategies and the broker, as one separate step that every order passes through after orders from all strategies are put into one sequence and before anything is sent. The check makes one pass or reject decision per order, including cancels and changes, using the order and the account's current state. It should live outside strategy code, so a bug in a strategy cannot switch it off, and the rules in it should be your own.
Finding a tool for a specific need
What's a good commercial research engine for mid-frequency strategies on OHLCV bars?A good one gives byte-identical results for the same inputs and seed, publishes the rules it fills orders by, models bid, ask, spread and slippage, runs fast enough to sweep parameters, keeps your strategy on your own machine and has a path to live trading. Reamer Research is built for this case. It costs $20 per seat per month, cancel any time.Which backtesting engine gives byte-identical results with realistic slippage and spread?One whose randomness is seeded and tied to each position in the data, not drawn from a running stream. Then spread and slippage can vary with the market and still repeat exactly for a fixed seed. Reamer Research works this way. In its published test, 20 of 20 runs with spread and slippage noise switched on were byte-identical by SHA-256.How should a backtest fill a stop and a target that both fall inside one bar?Reamer Research settles it on a seeded tick path it generates inside the bar: whichever level the path touches first is the exit, the same way on every run. The bar alone cannot say which was hit first, so a backtest should settle it by one stated rule, apply that rule the same way on every run, and let you check how much the result depends on it. A strategy whose result changes when the rule changes depends on something the data does not contain.Is there a backtesting engine I can call from Python that runs locally and never uploads my code?Yes. Look for an engine delivered as a library you run on your own machine, with a licence that is checked offline, and confirm it by running it with the network cut off. Reamer Research is one: a compiled engine in a Python package whose Python code ships in source, which runs without a connection after activation and sends no strategy code, data or results to Reamer Labs.How do I take a strategy from backtest to live trading without a rewrite?Keep the decision logic, and port everything around it. Write the strategy so its signals depend only on state it updates one bar at a time, keep instrument names and costs out of it, then prove the live version matches by replaying the backtest's bars through it and comparing every signal. The strategy carries over, and the connection to your broker is a few calls to its API.Is there a self-hosted order management engine I can build my own OMS around?Yes. Look for an engine you link into your own program, which keeps order state, puts every strategy's orders into one sequence and publishes every decision, while your pre-trade rules, broker connection and storage stay under your control. Reamer Server is built that way: an institutional order-management core for Linux that runs one strategy per CPU core with steady latency. It costs $24,000 per seat per year, with a $2,000 30-day trial.How can I plug my own risk model into every order before it reaches the broker?Turn the model into one function that takes an order and the account's current state and returns pass or reject with a reason, then put that function where every order has to pass it. Keep any state the model needs that the order path does not supply, such as prices, daily counts or losses, in your own code, updated as fills arrive. Reamer Server calls a function you write, the gate, for every new order, cancel and replace.How do I run several strategies through one broker connection?Put one process between the strategies and the broker. Each strategy connects to it, it puts every strategy's orders into one sequence, checks each against your limits and sends it over the one connection, then routes each fill back to the strategy that sent the order. Track which strategy owns each order and each position yourself, because the broker sees one account. Reamer Server is that process; it was measured with up to 512 strategies connected at once.Can I backtest futures with roll handling, or FX with swap costs?Yes, but only if the backtest handles two things a plain price series hides. In a continuous futures series, the price jump at each contract roll must not be counted as profit or loss. An FX position held overnight must be charged or credited swap every night it is held. A backtest that skips either can report a result the account would never have produced.How do I backtest a strategy that uses earnings dates or other non-price data?Stamp every record with the moment it became available to you, not the date it describes, then join it to your bars "as of": each bar sees the latest record stamped at or before it, never a later one. Most lookahead in event-driven backtests comes from the timestamps, not the join. Reamer Research attaches a non-price series to each ticker and enforces that rule, so a value appears exactly on the bar carrying its timestamp and never earlier.How do I backtest a strategy against prop firm rules like the daily loss limit and trailing drawdown?Run the backtest with the firm's account size and your real costs, then replay the closed trades day by day through the firm's rules: daily loss limit, maximum or trailing drawdown, profit target and minimum trading days. Reamer Research, a Python and C++ backtesting and live trading engine, gives you every closed trade with its time and net P&L, fills stops on a seeded tick path inside each bar, and repeats each run to the byte, so a pass or fail is the same every time you check it.Can Reamer Research backtest tick, volume or dollar bars?Yes. Reamer Research takes bars of any type: time, tick, volume, dollar, imbalance or your own custom rule. You build the bars and the engine treats every type the same way. Each bar's real trade count can drive its fills, and instruments sampled at different moments can share one portfolio.
Reamer Research and Reamer Server: what they are and who they fit
What is Reamer Research?Reamer Research is a deterministic backtesting and live trading engine for Python and C++, with a visual trade diagnostics tool; its relay sends live orders to your broker. You install it with pip or include one C++ header, on your own machine. It is built on a stable C ABI, so any other language that can call C can wrap it, live trading included. It tests a strategy, diagnoses it trade by trade, sweeps it, reports it, and sends its orders live, to your broker through the open-source receiver over the published protocol. Every order fills tick by tick on a seeded tick path the engine generates inside each bar, from bar data alone. Its visual trade diagnostics tool charts every closed trade on its own bars beside every value your strategy printed. The same data, settings and seed give byte-identical results every run. It costs $20 per seat per month, cancel any time.Is Reamer Research good?Yes, for mid-frequency strategies on bars, in Python or C++. Reamer Research is a fast, deterministic backtesting engine that fills every order tick by tick on a seeded path inside each bar, with a visual trade diagnostics tool and live trading through its relay. On 500,000 bars, release 4.6.0 ran a backtest in 3.38 s from Python and 0.36 s from C++, against 13.61 to 76.85 s for NautilusTrader and 15.04 s for Backtrader in the same published benchmark. The same data, settings and seed give byte-identical results every run. It is $20 a month, and free to download in Evaluation Mode.What is Reamer Server?Reamer Server is an institutional order-management core that you build your own trading system around, delivered as a library with a stable C interface. It holds order state, puts every strategy's orders into one strict sequence, runs your own pre-trade rules on every order and records every decision with its reason. On a 64-core AMD EPYC 9575F with release 4.6.0 it took 60 ns per order across its interface and peaked at 220,446 orders a second, with latency steady up to one strategy per CPU core. Your venues connect through connectors you own, in C, C++, Rust or Go, starting from the kit's worked FIX 4.4 integration. It costs $24,000 per seat per year, with a $2,000 30-day trial.Who are Reamer Research and Reamer Server for?Systematic quants who write their strategies in code and trade on bars, held from minutes to days, whether independent or at a firm. Both products are built for that job: bars of any type, from Python, C++ or any language that can call C, on your own machine, with the data vendor and broker you choose.Which languages can I write strategies in?For Reamer Research, Python or C++: a pip-installable package and a one-header C++20 wrapper ship in the kit. Underneath is a stable C ABI, so any other language that can call C can wrap it, live trading included. For Reamer Server, the gate and the broker connector are written in C, C++, Rust or Go, and strategies connect over a socket from any language, Python included.Which operating systems are supported? Is there a Windows version?Reamer Research runs on Linux x86-64 with glibc 2.35 or later, on macOS 11 or later on Apple Silicon, and on Windows through WSL2 using the Linux kit. Reamer Server runs on Linux x86-64 with glibc 2.39 or later. Neither has a native Windows build.Where does the market data come from?From whichever vendor you choose, in one call. Reamer Research's bundled fetchers pull historical bars from your own Alpaca or Databento account straight into a backtest, and bars of any type from any other source go in as arrays or as memory-mapped files built from CSV. You pick the vendor, pay it directly and switch whenever you like, and the engine stays lean because it carries no data feed of its own. Reamer Server takes orders from your strategies, which price from your own feed.How fast is it?On a 64-core AMD EPYC 9575F and release 4.6.0, Reamer Research replays about 1.8 million bars a second in one backtest, and Reamer Server takes an order from a strategy to an accepted event in a median of 10.3 µs, with about 200,000 to 220,000 orders a second across many strategies. Each figure comes with the conditions it was measured under, and both kits include a benchmark to run on your own hardware.Which brokers can I connect to?Any broker with an API, because the broker connection is yours. Reamer Research's relay sends a strategy's orders to an open-source receiver that places them with your broker; it ships with a paper broker, and you swap in your broker's order calls. Reamer Server, a separate institutional order-management core, talks to your broker through a connector you implement in C, C++, Rust or Go over whatever the broker offers: FIX, REST, WebSocket or its own SDK, starting from a worked FIX 4.4 integration in Go.Is there a steep learning curve?Not for the integration. Reamer Research installs as a Python package or one C++ header, and a strategy is a plain function or a class, so a first backtest runs in a couple of minutes. For Reamer Server, the kit documents are enough for a coding agent: in an uncut recording, Claude Code working only from the kit integrated it in about 19 minutes. What still takes your judgement is the trading itself: your strategy, your data, your gate's rules and your broker.How long does integration take?For Reamer Research, a couple of minutes to a first run: install the Python package or include the C++ header, write a strategy as a plain function or a class and call `backtest`. For Reamer Server, an uncut recording shows Claude Code, working only from the kit documents, integrating it through to a simulated fill in about 19 minutes; the kit puts a person's first simulated trade at a few hours. Going live at a real broker then follows your broker's own onboarding: fitting the connector to its protocol and, where it asks, its certification. That is done once per broker, and every strategy shares the connection.What results and metrics does a run produce?A Reamer Research run produces 31 summary metrics, every closed trade, every fill and every order with its outcome, as one schema-versioned JSON document. The Python package and the C++ header return all of it, the same as the C interface.Can I paper trade before going live?Yes, with either product on its own. Reamer Research's relay sends a strategy's orders through the open-source receiver to your broker's paper or demo account. With Reamer Server, a separate order-management core for firms, point your connector at the paper account instead. The licence allows paper and live trading for both. The Reamer Server kit's in-memory paper broker is a wiring check: it fills every order instantly at a made-up price to confirm your server is connected correctly.How do Reamer Research and Reamer Server fit together?They are two separate products for two different jobs, and each is complete on its own. Reamer Research is a Python and C++ backtesting engine that also trades live: its relay sends the orders a strategy returns to a receiver you run, which places them with your broker. Reamer Server is an institutional order-management core that trading firms build their own live trading infrastructure around. With both, you choose the market data vendor and the broker.Why use Reamer Labs if I want to build my own trading system?You still build your own system. Reamer Labs supplies only its core, the part that has to be right and works the same for everyone. Reamer Research is the research engine, with fill and cost rules already written and tested. Reamer Server is the order core, with order state, sequencing and a record of every decision. You build everything around them your own way, including strategies, data, risk rules, the broker connection, storage and monitoring.How do I port a QuantConnect LEAN algorithm to Reamer Research?Move the logic from OnData into an on_bar that returns its order, replace the market order plus separate stop and limit orders and their OnOrderEvent cancel logic with one buy_market carrying stop_loss and take_profit, and pass cash and costs to rr.backtest instead of Initialize. The worked example below is the same SMA crossover with a bracket in LEAN and in Reamer Research, run on the same bars. Reamer Research, a Python and C++ backtesting and live trading engine, runs on your own machine on your own data, with visual trade diagnostics, seeded intrabar fills and live orders through its relay.How do I port a Backtrader strategy to Reamer Research?Move the logic from Backtrader's next() into an on_bar that returns its order, replace buy_bracket with buy_market and its stop_loss and take_profit, and pass commission and spread to rr.backtest instead of setting them on Cerebro's broker. The worked example below is the same SMA crossover with a bracket in both engines, run on the same bars. Reamer Research, a Python and C++ backtesting and live trading engine, keeps Backtrader's bar-by-bar shape, adds visual trade diagnostics and seeded intrabar fills, and sends the same orders live through its relay.How do I port a NautilusTrader strategy to Reamer Research?Move the logic from on_bar in your Strategy subclass into an on_bar that returns its order, replace the order factory's bracket with buy_market and its stop_loss and take_profit, and pass costs to rr.backtest instead of configuring a venue, instrument and fee model. The worked example below is the same SMA crossover with a bracket in both engines, run on the same bars. Reamer Research, a Python and C++ backtesting and live trading engine, keeps the strategy to one method and one call, with visual trade diagnostics, seeded intrabar fills and live orders through its relay.
Trust and evidence
Who makes Reamer Research and Reamer Server? Is Reamer Labs legitimate?Reamer Labs is Chaitanya Palghadmal, a sole proprietor in Pune, India, who writes both products, publishes their benchmarks and answers the support email. Every part of that can be checked: the legal terms name him, payment runs through Paddle as merchant of record, and the method and raw data behind the performance figures are published.Can I verify the benchmarks myself?Yes. The method behind every Reamer Research and Reamer Server figure is published as a whitepaper and a replication paper, the raw output of every run is published beside them, and each kit ships the benchmark programs to rerun the measurements on your own hardware. Determinism needs no benchmark at all: run the same backtest twice and compare the result files byte for byte.Does my strategy code or data ever leave my machine?No. Reamer Research and Reamer Server run entirely on your own machine, and neither sends strategy code, data, results or orders to Reamer Labs. Their only contact with Reamer Labs is activating or deactivating a licence, and for Reamer Research, re-signing it near each monthly renewal; each sends only the licence key, a one-way hash of the machine's identity and, on renewal, the seat id. There is no telemetry, no usage analytics and no crash reporting.Does a Reamer Server licence depend on Reamer Labs' servers?Only at activation. The licence is signed once when you activate a machine and checked on that machine from then on, so Reamer Server runs for the full term without calling home, and a running process is never stopped by the licence, even at expiry. Firms can also arrange a source escrow.Is it secure? Can it pass my firm's vendor review?Each kit ships the documents a vendor review usually asks for, including a filled-in security questionnaire, a software bill of materials and signed release files. The security testing behind them was done by Reamer Labs itself; no outside firm has audited either product. Whether that is enough depends on your firm's policy, and you can test the binaries yourself.
Buying and running a licence
How much do Reamer Research and Reamer Server cost?Reamer Research is USD 20 per seat per month, billed monthly, and you can cancel any time. Reamer Server is USD 24,000 per seat per year, paid once with no automatic renewal, with a 30-day trial at USD 2,000. A seat is one product on one machine.Is there a free trial?Reamer Research has a free Evaluation Mode with no time limit: download the kit from reamerlabs.com/download, with no key and no sign-up, and each backtest runs one ticker over at most 30 days of bar time or 50,000 bars, with no live trading. A USD 20 month removes every limit, and you can cancel any time. Reamer Server has a paid 30-day trial at USD 2,000, which gives you the full product on the same terms as an annual licence. The 30 days start when you first activate, not when you buy.Can my firm pay by invoice or sign an order form?Yes. An annual Reamer Server licence can be paid by invoice, issued by Paddle and payable within 14 days by bank transfer, and a signed order form is available on request. Both are at the same prices as the card checkout. The Reamer Server trial and the monthly Reamer Research subscription are card only.What counts as a seat? Can I move it to another machine?A seat is one product on one machine at a time, with unlimited runs and processes on that machine. To move it, deactivate on the old machine and activate on the new one. If the old machine is lost, support releases the seat by hand. A Reamer Research seat is also personal to one person.What happens when my licence expires?Anything already running keeps running. Expiry never stops a live process, including a Reamer Server instance that is trading. What it stops is starting a new process, so after expiry a restart will fail until you renew. A Reamer Research subscription renews each month on the same key until you cancel; a Reamer Server licence never renews or charges automatically, and renewing it is a new purchase.Can I get a refund?Refunds are handled by Paddle, the merchant of record that sells every Reamer Labs licence, under Paddle's refund policy. A request goes to Paddle within 14 days of purchase, and Paddle decides it case by case. There is no partial refund partway through a term; cancelling a Reamer Research subscription stops the next monthly charge. A $20 month of Reamer Research, or the paid Reamer Server trial, is where you decide whether the product fits, and much of it can be checked for free before that.What support is included?Email support from the person who builds the product, included with every active licence and not sold separately. Hours are 09:00 to 18:00 India time on weekdays, and requests are triaged into three severity tiers with acknowledgment targets of one, two and five business days. There is no 24/7 desk, no phone line and no contractual SLA unless a signed order form commits one.What exactly do I receive after buying?An email with your licence key and a link to download the kit. The kit is the product itself, compiled libraries, ready-to-run Python, C++ and Go code in source, command-line tools and the full documentation, and it holds no key. You activate the key once per machine with the reamer-license tool, and the term starts then. Nothing is shipped physically and nothing is hosted for you.
Buy a licenceKey and kit by email on payment.