Backtester: model parameters and limitations
Backtester is modeling trades on historical data. To compare the calculations, keep not only the results, but also the exchange, the period, the regime, the commission and the settings of the indicators.
Additional VIP modes
When you buy a PRO for a year, you can use VIP Backtester tools: multi-mode, grid optimizer, hedge test and accelerated calculation. The payment terms are described in the articles Free, PRO and VIP.
The multi-mode checks one strategy for a selected list of assets. For each asset, a separate result is formed.
The test with the hash uses the robot's hash settings block (classic mode). It sets the parameters and levels of the classical hedge for historical calculation. This unit does not reproduce the aggressive mode of a hedge robot.
The Accelerated calculation is switched on by the switch x200 to the speed of calculation (VIP). The inscription indicates the mode, not the guaranteed acceleration of each calculation. When the hedge is turned on, the accelerated mode is turned off and becomes inaccessible.
A grid optimizer is used to select the combination of parameters. — The grid optimizer Its cloud launch requires a separate Modal account; the computing costs are related to that account.
Selecting and checking the list of tools is discussed in the article Multivariate: Checking strategies on multiple assets. — Multivariate: checking the strategy on several assets
The setting of the levels and limitations of the opposite position is described in the article Classic hedge in Backtester. — Classic hedge in the Backtester
Data and time frames
The step-by-step filling of the form is described in the article Backtester: First Calculation. — Backtester: first count Here are explanations for the parameters and the calculation assumptions.
The exchange you choose determines the source of the history. The calculation uses historical trades combined into price bars. The available period depends on the specific tool and the availability of data: the date of listing does not guarantee a continuous history to date.
The timeframe of the indicator is set in its card if that indicator has such a parameter. It determines which candles the signal is built on. This is a separate setting: changing the timeframe doesn't change the selected dates.
After the calculation, check the actual dates and the completion message. Missed data or premature stopping can shorten the time period covered.
Deposit, first order and leverage
Deposit the model's initial balance. He's not filling out an exchange account.
The First order specifies the volume of the quoted currency. It's not a pledge. The margin also depends on the leverage. For a limit grid, the calculation checks the funds for the first and the insurable orders together; when the signal is averaged, the additional margin is checked as the volume is added.
The Leverage and margin mode are involved in the calculation of available volume and liquidation. The permissible values depend on the limitations of the instrument. The leverage increase does not increase the model deposit itself.
The price and quantity are rounded by the instrument's limitations. Therefore, the calculated volume may differ from the input, and the order values too close to the order will be the same after rounding.
Commission and funding
Please indicate interest rates in the fields of the Comm. limit order and the Comm. market order. These are the entry parameters of the model: the values in the form do not confirm your actual exchange rate.
The calculation uses the market commission to signal market entry, average and exit. The limitation fee is applied to the limitation, the tax rate and the rate. Last Backtester's admission: a real stop on the exchange can be executed with another commission.
In the usual calculation periodic charges and write-offs of funding are not taken into account. The funding indicators are not available for adding to the form. So the result cannot be considered a complete reproduction of all the costs and revenues of a real futures account.
Net, profit and stop
The number of the S.O. sets the number of insurance orders after the first entry. With a limiting averaging, The SCO sets the initial step, the SCO multiplier of the next step, and the martinggel multiplier of the number in the next order. After changing the fields, check the net and the required volume.
The take profit is calculated from the average price of the position and reconstructed after the average. The null decomposes the corresponding exit.
The deviation limits the usual closing on the signal to the average price of the position. The value −1 disables this check. An indicator that requires a trade to be completed by force can close it without verifying deviation. It's a price threshold, not a guarantee of results after commissions.
Break-even stop
A Bybit calculation can use a protective stop based on the current average position price. Trigger must be greater than 0; level must be at least 0 and strictly lower than trigger. Both values are required, and averaging recalculates them from the new average.
The regular stop is measured from the first order and does not follow the average price. A 0% BE level is valid but does not guarantee a zero result because of fees and execution. See Break-even stop for full behavior and compatibility.
Balance mode of work
Without a statement, the result of the transactions is left in the model balance. Other options model the extraction of a positive result by the chosen rule. They affect the funds available for subsequent transactions.
In the monthly version, the check is performed before the new entry after the change of month: the excess over the initial capital is transferred to the model withdrawal. This is not a bank or exchange transaction or an automatic withdrawal schedule.
Why is the result different from the stock market?
The model checks the price performance of the bars. It doesn't reproduce the order of orders, the actual depth of the order book, network delays and actual slipping. If several levels are reached inside a bar, the model processing procedure is applied; the entire price path inside the bar is unknown.
The liquidation is calculated by the parameters of the model and the constraints of the instrument. This is not an exact historical reconstruction of all the rules of the exchange, the changes in rates and the status of your account.
The position open at the end of the period is shown separately and valued at the last available price. The end of the period doesn't mean the position was closed. Check it with the statistics on the completed transactions. The results on the history don't guarantee the results in real trading.