Backtesting Algo Tables

Logo	Name	Date	Instrument	TimeFrame	# Trades	% Profitable	Profit	Settings	Links	Comments
	EMA Algo Trader	6/12-6/23 10-3:30	NQ - 1contract	BR30	25	48	+6200	EMAs 8/34, No filter, Only 34B, Parabolic Stop 60,	Link1 Link2 Link3 Set	SL is not above previous bar high/low
	EMA Algo Trader	6/12-6/23 10-3:30	MNQ -5 contrats	BR30	50	36%	+2142	EMAs 8/34, No filter, Only 34B, Parabolic Stop 60,	link1 Link3
	EMA Algo Trader	6/12-6/23 10-3:30	MNQ -5 contrats	BR30	66	70%	+223	EMAs 8/34, No filter, Only 34B. Dynamic SLTP ATR 1	Link1 Link3	Test was done dynamically and not on bar based
	EMA Algo Trader	4/17-6/12 10-3:30	NQ -1 contract	BR30	160	33	+3750	EMAs 8/34, No filter, Only 34B. Parabolic Stop 50	link Link2 Link3 Link4	Note 2 months. Max drawdown is -1080
	EMA Algo Trader	4/142-6/12 10-3:30	NQ -1 contract	BR30	51	57%	+4590	EMA 21/50, dynamic SL, trail 60.	Link1 Link2 Link3
	EMA Algo Trader	4/142-6/12 10-3:30	NQ -1 contract	1000t	84	38%	-3255	EMA 21/50, dynamic SLTP ATR1	Link1 Link2 Link3	MOstly designed to run as 1:1 RR
	EMA Algo Trader	4/142-6/12 10-3:30	NQ -1 contract		144	47	-1880	EMA 8/34, dynamic SLTP ATR1	Link1	Usual settings
MNQ generates twice as much signals on the same parameters. The larger TF the less filter is becoming needed.BetterRenko bars perform better then Ninzarenko. When using dynamic one is betting on "this"accurate trade, when using parabolic one is betting on "overall" trade performace. Test5 suggests that there can be 2-3 days in drawdown before it will make a big leap and recover losses.

Name	Date	Instrument	TimeFrame	# Trades	% Profitable	Profit	Settings	Links	Comments
Stoch Algo Trader	6/12-6/23	NQ - 1contract	800t	18	61	+495	EMA Filter(50,100)Stoch(3,14),MinDist2, Stochlevels(85,15). No TimeFilter	Link1 Link2 Link3
Stoch Algo Trader	6/12-6/23	NQ contrats	500t	64	70	+2545	EMA Filter(50,100)Stoch(3,14),MinDist2, Stochlevels(85,15). No TimeFilter	link1 Link3 Link4
Stoch Algo Trader	6/12-6/23	ES 1 contrats	2000t	46	61	+675	EMA Filter(50,100)Stoch(3,14),MinDist1, Stochlevels(80,20),No Timefilter	Link1 Link2	Note ES
Stoch Algo Trader	6/12-6/23	ES 1 contrats	1300t	23	69	+662	EMA Filter(50,100)Stoch(3,14),MinDist1, Stochlevels(85,15),No Timefilter	Link1 Link2
Stoch Algo Trader	4/17-6/12 10-3:30	NQ 1	2000t	8	75	+680		Link1 Link2

Name	Date	Instrument	TimeFrame	# Trades	% Profitable	Profit	Settings	Links
Power Algo (BOP) Trader	4/7-6/12 10-3:30	NQ - 1contract	800t	198	54%	+890	Engulf Only	Link1 Link2 Link3 Set
Power Algo (BOP) Trader	6/12-6/23	NQ 1 contrats	800t	30	55%	+55	Only engulf, No time filter	link1 Link3
Power Algo (BOP) Trader	6/12-6/23	NQ 1 contrats	500t	111	53%	+265	only jump	Link1 Link3
Power Algo (BOP) Trader	6/12-6/23	NQ -1 contract	800t	160	14%	-1400	Jump (child slave) No Timefilter	link Link2 Link3
Power Algo (BOP) Trader	6/12-6/23 10-3:30	NQ -1 contract	800t	14	57%	+395	Cross (swaying allowed)	set1 Link2

Name	Date	Instrument	TimeFrame	# Trades	% Profitable	Profit	Settings	Links
PSAR Algo Trader	4/7-6/12 10-3:30	NQ - 1contract	800t	263	41%	+6500		Link1 Link2 Link3 Set Link5
PSAR Algo Trader	6/12-6/23 10-3:30	NQ 1 contrats	800t	223	43%	+5250	EMA 100 Filter	link1

Logo	Name	Date	Instrument	TimeFrame	# Trades	% Profitable	Profit	Settings	Links
	Bounce Trader Algo Trader	4/7-6/12 10-3:30	NQ - 1contract	NR20/10	252	69%	+645	Only 34B	Link1 Link2 Link3
	Bounce Trader Algo Trader	4/7-6/12 10-3:30	NQ 1 contrats	NR30/15	141	67%	-20	Only 34B	link1 Link2



bounce trading system is original first bounce trader algo that is based on Ninzarenko bars

Logo	Name	Date	Instrument	TimeFrame	# Trades	% Profitable	Profit	Settings	Links
	Trend Magic Algo	6/12-6/30 10-3:30	MNQ - 1contract	800 t	77	55%	-220	Only 34B	Link1 Link2




bounce trading system is original first bounce trader algo that is based on Ninzarenko bars

Note: if there is No time selected it means no time filter is used and strategy trades around the clock which in most cases not practical. See links to find out what hours were mostly profitable and what not.

Note: 2 weeks tests run about 20 mins and they can be performed relatively fast. Longer time tests such as 2 months takes about 5 hours on VPS so those tests are fewer. All tests run by market replay.

Note: Tests run on different machine maybe different. Also it maybe that while tests ran it used unsynced or skewed data. To make sure tests run accurately we need to be rerun with refreshing data and ninjascript. These tests were not rerun for double check. Data used Apex

Note: If tests are negative it means that optimal settings were not found for that period. It doesnt nessessary disqualifies the strategy.

BR – BetterRenko bartype, NR – Ninzarenko bartype, 800t – 800 ticks

All tests are performed using market replay and data from Apex. When special settings are used its noted in comments section. First we want to run 2 week tests sample to see what has potential to be profitable, then we run 2 months test for confirmation.

Some algos are so versatile that it would take months to test them out fully.

Login

Signup