Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
Report Generator Tests / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
* Market orders wait for fresh data instead of filling on stale prices
A market order would previously fill immediately on the most recent
available data even when that data was older than StalePriceTimeSpan
(default one hour), only attaching a warning. This is unrealistic for a
coarse resolution asset (hour/daily) where the latest bar is the stale
previous close when the order is placed mid-bar or via an intraday
scheduled event.
The default fill models (FillModel, EquityFillModel, FutureFillModel) now
wait for fresh data instead of filling on a stale price, but only for hour
and daily resolutions; the order fills when the next bar closes. For
minute/second/tick subscriptions the previous behavior is kept (fill on
the stale price with a warning), since stale data there is a genuine gap
rather than a bar still forming.
Adds HourResolutionMarketOrderStalePriceRegressionAlgorithm, updates the
FillOutsideHours daily expectation, and regenerates statistics for the
hour/daily algorithms whose fills change. FutureOptionDaily buys and
liquidates a day apart now (a same-day buy + liquidate cannot fill on
daily data once stale fills are disabled).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Normalize and correct StalePriceTimeSpan XML docs
The interface and class docs now match and reflect the actual behavior:
the wait-for-fresh-data only applies to hour/daily resolutions, while
minute/second/tick subscriptions still fill on stale data with a warning.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Fill resting market orders at the bar open instead of the close
A hour/daily market order that was resting before the current bar opened
(it predates the bar - placed after the previous close or while waiting
for fresh data) now fills at the bar open, the price when trading resumed
(like a MarketOnOpen), instead of the bar close. Orders placed during the
bar still fill at the current/close price, so intraday mid-bar fills are
unchanged. Equity fills are unchanged (resting equity orders are already
converted to MarketOnOpen by QCAlgorithm.MarketOrder).
Adds the shared FillModel.GetMarketFillPrice helper used by the base
FillModel and FutureFillModel, a unit test, and regenerates statistics for
the affected daily/hour futures, index and crypto regression algorithms
(order counts unchanged, only fill prices).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Add regression algorithm asserting resting market orders fill at the bar open
RestingMarketOrderFillsAtBarOpenRegressionAlgorithm buys a daily future on the
bar that delivers it (fills at that bar's close) and submits a liquidation while
the market is closed (overnight pulse, no fresh bar). The liquidation rests and
fills on a later bar at the bar open, not its close - asserting the new
GetMarketFillPrice behavior. The in-bar buy is asserted to fill at the close, for
contrast.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Carry the bar start time on Prices instead of re-reading the cache
Add Prices.Time (the bar start, mirroring BaseData.Time/EndTime), populated from
the source bar/tick in every GetPrices path. GetMarketFillPrice now uses
prices.Time directly instead of a second asset.Cache.GetData() lookup. Behavior
is unchanged (prices.Time equals the previously read cache time).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Add regression algorithm asserting in-session hour orders fill at the latest close
HourMarketOrderFillsAtBarCloseRegressionAlgorithm submits an hour resolution
market order mid-bar (via an intraday scheduled event) while the market is open,
using the default one hour StalePriceTimeSpan. It asserts the order fills
immediately at the latest available bar's close - not waiting and not at the bar
open - since the latest bar is within the stale window. Guards the resting-order
open-fill behavior against affecting ordinary in-session fills.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Regenerate custom fill model algorithm statistics for the open-fill change
CustomModelsAlgorithm and CustomPartialFillModelAlgorithm subscribe SPY at hour
resolution and their custom fill models delegate to base.MarketFill, so resting
orders now fill at the bar open. Regenerate their statistics (C#/Python) and the
inline expected statistics of the PEP8StyleCustomModelsWork test.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
API Tests / build (push) Has been cancelled
Benchmarks / build (push) Has been cancelled
Report Generator Tests / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
Syntax Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
* Add SeedInitialPrices algorithm setting
This is true by default and indicates that the engine will seed initial prices right after the security is added or selected
* Update regression algorithms
* Update regression algorithms
* Update regression algorithms
* Refactor default securities seeding
* Minor fix
* Minro fixes
* Cleanup
* Updated and add regression algorithms
* Address peer review
* Centralize logic to get last known data for multiple securities
* Some cleanup
* Minor build fix
* Minor fixes
* More logic centralization
* Some more cleanup
* Cleanup
* Update regression algorithms and minor fixes
* Update regression algorithms
* Minor fix
* More minor fixes
* Update regression algorithms
* Cleanup
* Minor test fix
* Address peer review
* Minor fix and performance improvement
* Fix to seed open interest data
* Minor test fixes
* Address peer review
* Minor change
* Minor revert
* Minor fixes and improvements
* Disable initial seeding by default
* Minor fixes
* Cleanup
* Cleanup
* Minor fix
Build & Test Lean / build (push) Has been cancelled
Report Generator Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
Benchmarks / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
* Default daily precise end times
- Enable by default daily precise end times. Updating stats
- Minor fix for algorithm manager consolidator updates, adding new regression test
asserting behavior and updating others
- Minor fix for SubscriptionData creator avoid round down on warmup if
not appropiate
- Adjust consolidators to emit on daily strict end times if requested
daily resolution and setting enabled
- Updating regression algorithms
* Skip daily data on extended market hours
* Some cleanup and self review
* Revert unrequired change
* Dynamic spdb
* Minor changes and unit test
* Get MHDB and SPDB refresh period from config
* Add databases refresh period to algorithm settings
* Address peer review and add unit test
* Daily data Time & EndTime Improvement
- Adjust daily data Time & EndTime to actually reflect the time of the
data used, for example US Equity from 9.30 to 4PM. Adding new unit and
regression tests
* Refactor solution to use enumerator
- Refactor daily strict end times solution to be through enumerator
usage, so it applies for history providers too
* Minor fixes
* Revert fill forward enumerator change
- Revert FillForward enumerator causing stats changing, enhancing unit
tests
* Some cleanup
* Improve handling of live trading FF enumerator
- Improve handling of live trading FF enumerator, by adding support for
bars to arrive with a delay so we can handle auction close/option
prices or data providers which might have some delay making the data
available. Adding new unit tests asserting the behavior
* feat: new `TradingDaysPerYear` prop in IAlgorithmSettings
* feat: getTradingDayPerYear by BrokerageModel
feat: init algo tradingDaysPerYear in brokerage|BacktestingSetupHandler
refactor: nullable `TradingDayPerYear`
* remove: hardcoded value:252 in statistics
* test: `algorithm.Settings.TradingDaysPerYear` for different brokerage
fix: tests calculation of `AnnualPerformance()`
* fix: Report generator by tradingDayPerYear
* fix: hardcoded value in PortfolioStatisticsTests
* fix: tradingDayPerYear in AnnualPerformanceTest
remove: extra spacing
* feat: backwards compatibility for TradingDaysPerYear
* feat: several UTest for AlgoSetting's TradingDaysPerYear prop
* fix: existed algo + test
* feat: regression algo with checking of tradingDayPerYear in OnEndOfAlgorithm()
* refactor: GetBrokerageTradingDayPerYear -> SetBrokerage... in BaseSetupHandler
refactor: carry out tradingDayPerYear at the top level and pass variable
* remove: default valuine in test param
remove: extra comment.
* fix: missed init TradingDaysPerYear in report.portfolioAlgo
fix: hardcoded value to tradingDaysPerYear
* fix: nullable TradingDaysPerYear in AlgorithmConfiguration
feat: UTest for PortfolioLooperAlgo
feat: add advanced description
* feat: test PortfolioStatistics with different tradingDaysPerYear
* revert: PortfolioLooper with TradingDaysPerYear implementation
* revert: nullable TradingDaysPerYear in AlgoConfiguration
remove: SetBrokerageTradingDayPerYear in BaseResultHandler
* style: remove empty row
Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
* Deprecates IAlgorithmSettings.DataSubscriptionLimit
We will let the brokerage check and notify users that they have subscribed to datafeeds beyond their quota.
* Removes Unit Test
Also remove usage in `EmptyEquityAndOptions400Benchmark`.
Regression Tests / build (push) Has been cancelled
Benchmarks / build (push) Has been cancelled
Build & Test Lean / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
Python Virtual Environments / build (push) Has been cancelled
* Implement Trailing FreePortfolioValue
- Implement Trailing FreePortfolioValue by default, users will be able
to set it to a fixed number if desired. Adding regression algorithm
- Setting the default 'MinimumOrderMarginPortfolioPercentage' from 0 to
0.1% of the TPV to avoud tiny trades by default
* Update existing regression algorithms
* Address reviews
- Send warning message to the user if a trade does not happen due to the
default setting of the minimum order margin percentage value
* Address reivews
* Rename TotalPortfolioValueLessFreeBuffer
* Update new regression algorithm
Build & Test Lean / build (push) Has been cancelled
Regression Tests / build (push) Has been cancelled
Research Regression Tests / build (push) Has been cancelled
* Respect warmup resolution given
- The data feed will respect the warmup resolution given and override
the resolution used by the algorithm when adding a subscription.
Updating regression algorithm to keep previous statistics. Adding new
regression algorithm asserting the desired behavior
* Testing improvements
- Add more unit tests and regresion test
- Add missing data for crypto
- Fix bug with FFed data crossing after the end time of the warmup
request
* Add more Warmup resolution regression algorithms
- Adding more warmup resolution regression algorithms, using
Settings.WarmupResolution and an option selection case
* Add more warmup regression tests
- Adding more warmup regression tests.
- Will no longer skip universe selection subscriptions from warmup
resolution enforcement. Updating regression algorithms data points
* Fix bug with data rounding
- Fix data rounding bug when warmup resolution is set to a different
value than the original configuration. Updating regression algorithms
to assert the expected behavior
* Address reviews
- Revert regression algorithms changes to use Resolution during warmup.
Updating their stats.
- Adding new regression algorithms asserting the behavior warming up
using a timespan and no warmup resolution
- Fix bug where data used to warmup the 'normal' enumerator will make it
through into the warmup time span. Updating tests
* Address reviews
- Add missing comments, explaning warmup algorithms time span
calculations.
- Revert changes in existing `WarmupOptionTimeSpanRegressionAlgorithm`
to reduce diff to minimum
- Adding new warmup unit tests asseting algorithm warmup start time, for
different combinations of bar count, timespan, resolution
Regression Tests / build (push) Has been cancelled
Build & Test Lean / build (push) Has been cancelled
* Filter out small orders based on Setting
- BuyingPowerModel will filter out small orders based on algorithm
setting, a % of PTV, instead of hard coded 1 share value. Addin unit
and regression tests
- Updating regression algorithms to use new setting, reduce order trades
* Update regression algorithms
- Adding `FreePortfolioValue` to be set after algorithm initialize based
on the `TotalPortfolioValue` and the `FreePortfolioValuePercentage`
- Updating regression tests
- Adding new regression test
- Adding check for minimum order value at `BuyingPowerModel`
- Adding new `AlgorithmSettings` Min and Max absolute portfolio target
percentage
- Adding new `PortfolioConstructionModel.FilterInvalidInsightMagnitude()`
helper method that will be used by the `BlackLitterman` and
`MeanVariance` optiomization portfolio construction models to skip
insights with extreme magnitudes that will cause exceptions
- `PortfolioTarget.Percentage()` will now verify requested percent is
withing the settings values
The market fill warning introduced in #2869 was considering any price older than the fill time as stale (even if only a few seconds older), causing many false alerts.
This is being changed now by adding a new configurable algorithm setting: StalePriceTimeSpan (with a default value of one hour).
Martin-Molinero