robot trader prototype product research roundup tickerbase backbase candbase bull sector yahoo...
TRANSCRIPT
Robot Trader
Prototype Product Research
Roundup
TickerBase
BackBase
CandBase
BullSector
Yahoo
TadaFilter
UserScripts
OptionMaker
TzoneWebSite
6000+tickers
1000+tickers & open-int
1000+stock histories
50+candidate stocks
1:30pm
2:00pm
2:30pm
CandBase
Orders
5:30am
OXStrategy
Yahoo
6000+option chains
CandBase
OXOrders
OXPositions
Yahoo
EDWebGen4:30pm
EDWebAgents6:30pm
OpenMan
6:30am
User Shell Scripts
Perl Driver Scripts
TZ API Modules
External Websites
VPN Internal Website
Pgsql Databases
Tzone Trading Cycle & Simulation Architecture
150+candidate options
Review of 150+candidate options Selection of 20+
candidate trades
Simulation Control
Replay and Backtesting
Selection List
Intraday Stock &Option Chains OptionsFeed
6:30amStockBarsFeed6:30am
CandBase
DayTrader6:30am
(Remote Procedures)
OX Agent
Broker
TZSimWebSite
BackBase
SimulatorAnytime
TZSim Modules
TZSim Website
Manual Approval Meeting Realtime
NonSim
NonSim
Phase 1 Prototype
“Tada” Theory• Technical Analysis of Daily Action (Tada - Holon)
– Stock History (starting 6 months prior to trade date)– Selection - GT Buy/Sell signal on trade date– Tada Assemblies (Bottom up Hierarchies – Holons)
• Various Indicators (MACD, MFI,RSI)• Various Signals (Indicator Combinations, eg. MACD-MFI)• Various Vanes (Buy/Sell Decision Makers – GT “systems”)
• Tada 13 Theory – ZigZag Swing Trading– Focus on Stock Cycle Channel (Range)– De-trended Channel for Trending Stocks– Re-trade on closing to catch swing
• Confirm with Rocket Science (Ehlers)– Hilbert Frequency Domain Analysis (Future)
ZigZag Swing Theory (early version without de-trending)
Average of bottom peaks
Average down angleAverage of top peaks
EDT*
* EDT: Estimated Days to Target
BTF-WID (Weekly-Intraday)
Back to the future (Daily)
BTF-D
Bullish-Bearish Degrees
• Wide Trades – Coded ‘B’ or ‘S’ (BUY or SELL)– ‘B’ – Below average-bottom line on upslope or up day
– ‘S’ – Above average-top line on downslope or down day
• Edge Trades – Coded ‘b’ or ‘s’ (buy or sell)– ‘b’ – At or near-above average-bottom line, etc.
– ‘s’ – At or near-below average-top line, etc.
• Mid Trades – Coded ‘bs’ or ‘sb’ (buy-sell or sell-buy)– ‘bs’ – Below mid-point between avg-bottom & avg-top
– ‘sb’ – Above mid-point between avg-bottom & avg-top
Filtering with Tandem Tadas
Tada Rating: GainRate = TradeGain * PeakRate ($/month) ($/swing) (swings/month)
Merging of tandem tada
ratingsAveraging of 2 tandem tadasDifferent according to gainparameter of ZigZag Indicator
Edge trade (s)
Wide trade (S)
Wide trade (B)
CandBase
Tadamon Daemon
Tada15 50+candidate stocks
Orders
TradeBase
Simulation Architecture Re-factoring
OpenMan
Selection List
EDWebAgents
EDWebGen
CandBase
BackBaseOXBroker
Clock Simulator
TzoneWebSite
Review of 150+candidate options Selection of 20+
candidate trades
Manual Approval Meeting
Simulation Control
Replay and Backtesting
TZSimWebSite
DayTrader
Intraday Profiles
HistBase
Roundup
TickerBase
16000+tickers
1000+tickers & open-int
OptionMaker3
CandBase
150+candidate options
BackBase
TadaFilter4 Spliced Intraday histories
Updated Daily—Comprehensive daily data up to yesterday
Tada14
TadaFleshout
Tada Daily Price histories 2000+
option chains
CandBase
TadaFleshout
Tada Daily Price histories
Machine-Gun & Re-Order Requests
TradeBase
Machine-Gun (MG) Simulation Architecture
EDWebGen
OXBroker
Clock SimulatorDaily Pause Triggering
(Semi-Automatic Firing)
Interactive Simulation Control
Replay and Backtesting
TZSimWebSite
DayTrader
HistBase
Roundup
TickerBase
16000+tickers
1000+tickers & open-int
2000+option chains
BackBase
Tadamon
Tada15
Tada14
EDWebAgents
Orders
30sec-Day Reply Cycle
30sec-Day Request Cycle
CandBase
TickerBase
TadaJobTada14
Tada15
OpenMan
BrokerSimulator
Phase 2 Prototype
Simulator Run Modes & Objectives
• Single-Shot Shotgun Mode (Each pellet is a trade)– Tracks the trajectory of one-day’s trade launch
– Re-trades on swings qualified by TadaMon daemon• Request-reply cycle through TickerBase feeds work from DayTrader
• Current positions only, TadaMon generates new trades
• Machine-Gun Mode – Fires Shotgun each day– New OpenMan Cycle for each day (new batch of trades into the mix)
– Deals with aggregate of all trades (easily 500-1000 active trades)
• Primary Objective – Test & Validate DayTrader– Single shot shotgun (now) can test position management
– Machine shotgun needed to validate portfolio management.
• Secondary Objective – Optimize Trading Strategy– Single shot is adequate for this, mostly
• Current Status – Machine-Gun Not Yet Operational
Re-Trading Simulations• Next slides illustrate re-trading simulation
– Used to validate TadaMon daemon
• Re-trading focused on Puts & Calls only– A move away from combination trades (e.g. spreads)
– But Straddles/Strangles (mid-trades) were openers• According to a naïve pairing strategy, subsequently discarded
• TadaMon results initially accepted for wide date intervals– Exact date of closing trade might not have new opener
– Any TadaMon opener within -2 days to +5 days was accepted• To check out the re-trading logic
– Only some closings could catch a new opener• Wide intervals accepted less than optimum re-trades
Retrade
Retrade
Retrade
What Has Been Achieved
• Development of Fully Automatic Trading Software– Proof of automatic factory trading of options
• Reliable sifting of best trades each day– From all stocks in US markets– Optimization of option trades for those stocks
• Automatic opening of best trades through broker website• Automatic position management through broker website
– Automatic day-trading with intraday feeds– Automatic re-trading after profit-taking
– Proof of swing trading factory strategy• Automatic selection of best batch of trades daily• Continuous profit taking to maximize revenue
• A Basis for revolutionary wealth generation
Work Still to be Done
• Hilbert Tadas – To Confirm ZigZag Tadas– Available in TA-LIB, a C-library
• We converted to a Perl Extension – not yet applied
– Important for TadaMon Dynamic Control
• Implementation of Dynamic Control– TadaMon Daily Updates of Active Positions
• Check if trades behaving as forecast
• If not, switch to exit lane strategies
– Strip Mining – Multi-Lane Closing Strategies• Dynamic Lane Changing is Control Mode
Strip Mining
• Fast Lane – Highest Rated Swingers– High GainRate, High Peakrate– High pumpers, up-stroke & down-stroke– High threshold trailing stop strategies
• Slow Lane – Change-Up Situations– Not Meeting expectations, but not losers– Low theshold trailing stop strategies
• Exit Lane – Suspected Losers– Minimum loss strategies:
• Timestops, TimeTrailStops, etc.
– Aim: Cut losses early
EDTBTF-D
Key to Strip Mining: TadaMon Monitoring of Positions
Check if EDT and ZigZag angle is accurate (+- 50%) If yes, set trailing stop when target is crossed If no, switch lanes to slow or exit lanes
Re-trade
Trade Date
TadaJob HistoricalTechnical Analysis Trade Selection
TadaMon ActivePosition Management Trade Selection