About our algos
- Each algo runs in a dedicated proprietary environment, serverside, not on clientside machines.
[Our algos are not scripts, cannot be downloaded or similar]. - Each algo is composed by many modules, dynamically configured to react to markets movements.
- Our algos are specialist, not just expert systems, covering: eurodollar, oil, gold, no other instruments.
- Our Algos are broker independent, bank independent: would not work in retail environments.
- In 15 years we developed a distinctive algo generation / development paradigm, which uniquely identifies any algo of ours from competitor’s ones.
[Paradigm is classified and eventually part of Patent and/or Trade Secret]. - Our algos are trained to defend from counterpart speculative behaviors, including counterpart malpractices profiling.
- Our algos often show sellside behavioral patterns: brokers or market makers algo-pools don’t allow it as would unmask their conflict of interest (they always maximize traded volume, our algos don’t).
- Our Algo enviroment is set not to improve profitability: in an ever evolving arena improving efficiency, remainin g profitable is hard enough, forcing a model to improve its profitability and efficiency just degrades risk quality.
- Our algos are trained to spot and exploit loopholes which are often kept hidden by exchanges, banks, brokers and especially market makers which often are wilfully engaged in masking massive streamline profit sipping from their customers in that way.
Example of Market Environment Situational Awareness:
Duringa a flash-crash customers orders get frozen, not netted: get suspended.
Brokers net them instead but Money does not “just disappear”
Brokers and banks never say it this way to their counterparts.
As by expertise , our algos are aware of such industry malpractices, often used as an excuse for brokers topull stops or net positions using a force majeure clause or an arbitrage clause, which are just disguised tools to mask a huge conflict of interest not related to real market dynamics.
Example of Expertise Driven Learning :
From raw intelligence to gain of function, evolved on the field.
None of your algos owns a “firewall against biased, lagged or clickbait prices”.
That’s because no broker would ever allow a customer to filter (mark against) his prices.
While our algos have a specific module just to counter just such brokerside behavior,
which is just one conflict of interests… among many.
Algo Modules
Algo Class | Market Side | Strategies | Engine | Algo Module details |
---|---|---|---|---|
Event Driven | Both | passive | Machine learning | Externalities: good / bad / expected / unexpected / rational / irrational. Panic news hits: on liquidity / on Governments / on markets. Panic news reaction: on liquidity / on Governments / on markets. Panic news aftermath: correlations / dampers / duration / envelope. Political Spin doctoring: exploiting / debunking / reverting / spotting. |
Liquidity Aware | Both | active | Machine learning | Asset pricing failure: biased / intentional / systematic / systemic. Bad: projections / expectations / forecasts. Hot money / cold money. Illiquid: markets / assets. Inconsistencies: margin system / collateral / underlier. Liquidity plunges / drains / surges. Margin failure: contingency repricing / force majeure / overleverage. Situational awareness failure. Smart money / dumb money. Sudden asset repricing / devaluation / guarantees expirations. Unexpected exposure: externalities / event mismanagement. Unfair competitor practices: hijacking / derailing / lagging / opportunistic disinfo. |
Trading Modules | Buyside | active | Specialist A.I. Expertise driven | Alternative pricing. Arbitrage. Hedging: full / partial / triangular. Market Profiling. Predator-prey: Fittest / Gause. Proprietary pricing model. Price filters. Scalping: LFT / HFT / LLHFT / ULLHFT. Sniffing. Trend Following. Volatility based. Volume based. |
Anti-Disruption | Both | active | Specialist A.I. Expertise driven | Anti-derailers. Anti-iceberg. Anti-frontrunning. Anti-hedging: full / partial / triangular. Anti-laggers. Anti-lowballing. Anti-mirroring. Anti-newstraders. Anti-order-pool. Anti-panic: mongers/riders. Anti-price:casters / stuffers. Anti-quote-injectors. Anti-sandboxing.Anti-scalping. Anti-slippage. Anti-sniffer. Anti-sniping. Anti-spread. Anti-spoofing |
Anti-Infiltration | Sell side | passive | Specialist A.I. Expertise driven | Counter-arbitraging. Counter-baiting. Counter-looping. Counter-masking. Counter-derailing. Counter-disrupting. Counter-picking. Counter-profiling. Counter-socialtrading. Counter-shredding. Counter-shuttering. Counter-skewing. Counter-spiking.Counter-stuttering. |
Collaborative | Sell side | active | Machine Learning Q learning | Network distributed Algo Pool. Swarm intelligence. Trial & error. |
Predatory | Buy side | active | Expertise driven specialist A.I. | Counter-Profiling: Pool/Dealer/Market. Deep Learning (Mimic): algo reverse engineering from Trade Reports. Profiling: Pool/Dealer/Market. |
Correlation Matrix | Buy side | passive | Machine Learning Q learning | AI driven dynamic correlations: direct, inverse, anti. . Dynamic pattern recognition, with anticipation and lag awareness from 2d Heatmaps to 4D Hyperspace cross matrix (market-wide). |
Allowed modules by User Class
Dark Pools Plunge Protection Teams Sovereign Funds | Commercial Banks Electric Power Producion & Distribution Companies Oil&Gas | Treasury Engines Liquitidy Engines Clearing houses | Invest. Banks Hedge Funds | Asset Management Companies Wealth management Companies Family Offices |
|
---|---|---|---|---|---|
Event Driven | Allowed | Allowed | Allowed | Allowed | Allowed |
Liquidity Aware | Allowed | Allowed | Allowed | Allowed | Allowed |
Trading Modules | Allowed | Allowed | Forbidden | Forbidden | Allowed |
Anti Disruption | Allowed | Forbidden | Allowed | Allowed | Forbidden |
Anti Infiltration | Allowed | Allowed | Allowed | Allowed | Forbidden |
Collaborative | Allowed | Allowed | Allowed | Forbidden | Forbidden |
Predatory | Forbidden | Forbidden | Allowed | Allowed | Allowed |
Algo Deployment by Users Class
Dark Pools Plunge Protection Teams Sovereign Funds | Commercial Banks Electric Power Producion & Distribution Companies Oil&Gas | Treasury Engines Liquitidy Engines Clearing houses | Invest. Banks Hedge Funds | Asset Management Companies Wealth management Companies Family Offices |
|
---|---|---|---|---|---|
User Acceptance Testing | Specific | Specific | Specific | Standard | Standard |
Algo Clusters: from deployment to production: | 1 month after signing NDA+NCA | 2 weeks after signing NDA+NCA | 1 month after signing NDA+NCA | 2 weeks after signing NDA+NCA | 48 hours after signing NDA+NCA |
Distribution: per World Time Zones | Colocated | Remote | Colocated | Remote | Remote |
Expected Volume [signals/day] | [restricted] | [restricted] | [restricted] | [restricted] | 40 |
Virtual Machines needed | [restricted] | [restricted] | [restricted] | [restricted] | 6 |
Algo Basket - Structure Example:
25 Algos are undisclosed [ NDA ]
75 Algos (regrouped by class) are listed below
Number of Algos for each class | Algo Class | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
25 | Undisclosed Algos [ NDA ] | ||||||||||||
3 | B-book counter-profiling | ||||||||||||
3 | Chaos driven | ||||||||||||
3 | Evolutive | ||||||||||||
3 | Event driven | ||||||||||||
3 | Distributed A.I. | ||||||||||||
3 | Gaming | ||||||||||||
3 | Hedging | ||||||||||||
3 | Market Profiling | ||||||||||||
3 | Mimic | ||||||||||||
3 | Opportunistic | ||||||||||||
3 | Panic driven | ||||||||||||
3 | Parassitic | ||||||||||||
3 | Position trimming | ||||||||||||
3 | Predator-prey | ||||||||||||
3 | Predatory | ||||||||||||
3 | Random walk | ||||||||||||
3 | Revolving | ||||||||||||
3 | Sniffers | ||||||||||||
3 | Swarm Intelligence | ||||||||||||
3 | Trader Counter-Profiling | ||||||||||||
3 | Trend following | ||||||||||||
3 | Triggered by Market News | ||||||||||||
6 | Order Recycling | ||||||||||||
3 | Vulture |
Algo Basket - Composition Example:
Dedicated Algos | Instrument | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
22 | EURUSD | |||||||||||||
22 | Oil | |||||||||||||
22 | Gold | |||||||||||||
34 | Exchange Indexes |
The Exchange Indexes partition: is a dynamic mix of major Exchange Indexes such as: S&P500 Dax30, VIX etc.
Algo raw output signals can be either long , short or flat and cannot be reconfigured locally .
Each single raw signal output is digitally signed to avoid man in the middle Basket Manipulation:
No "Picked ETFs"
No "Picked Stocks"
No "Picked Funds"
No "cryptos"
Proprietary Index Example:
Autoscalp Multi-Benchmark Index.
Specs:
Objectives: pure alpha seeker.
Raw Algo output (raw signals) to be transformed by counterpart in trading signals.
Risk brackets are in line with other Autoscalp algos: +2%,-4%.
Brief:
Composed of 40+ benckmark indexes, only tradable ones.
Iso-risk weighted, same weights on all components.
Dynamics:
Basket redesigns itself every 48 hours by a pool of autonomous proprietary algos.
Failsafe and Index Protection:
Index is self protecting against typical indexes flaws.
[ I.E. SP&500 flaw: 1% of its components move S&P500 90% of times ]
Many other forms of index flaws exists:
some are known some unknown some are even knowingly
unacknowledged by exchanges thus a specific proprietary module
takes care of fully exploiting this highly arbitrable grey area.
(Failsafe Sniffer Algo).
Index Neutrality:
The index refreshing process is completely automated and uneditable.
Every 48 hours all Multi-Benchmark Index components get replaced.
During the index refreshing process, process modifications are locked out
to avoid attempts to bias the index toward any direction by any possible stake holder.
Why neutrality and lockout is needed:
There is a whole sector of Bank Pseudo-advisors, Fund pseudo-advisors , even pseudo-robot advisors,
faking objective bias while are pushing their ETFs insetead.
Then the Locking-out of the modifications during Index rebalancing was a needed feature.
The Confusion Matrix monitoring our Machine Learning Process :
Confusion Matrix measures the performance of a model.
As samples belong to two classes: forecasts can be correct (if profits) or wrong (if loses).
As the classifier predicts for a given input sample,
computing the Confusion Matrix of 2.8+ million algorithms would be totally useless.
So what we do is keep the Confusion Matrix up to date given the algorithms running at a given time.
Apparently other variables to consider would kick in such as: size and money management etc.
but these are just applied techniques and methods to use the algos, not the algo themselves.
That assumed, the expected instantaneous Confusion matrix of the Algos in use are (rounded):
64 True Positives : The cases in which we predicted YES and the actual output was also YES.
81 True Negatives : The cases in which we predicted NO and the actual output was NO.
36 False Positives : The cases in which we predicted YES and the actual output was NO.
19 False Negatives : The cases in which we predicted NO and the actual output was YES.
which means:
64 correct signal: result was profit.
81 correct don’t trade signal: would have lost money if traded ignoring the don’t trade filter.
36 bad signal: result was a loss.
19 bad don’t trade signal: would have profited if traded ignoring the don’t trade filter.
Accuracy for the Matrix can be computed by taking average of the values lying across
the “main diagonal”. Although We feel is is more easily understandable as posted above.
Note:
While algos looks over-specialized on “don’t trade” signals, that just depends from Expertise:
Forecasting the perfect moment and price to enter a deal is difficult
while expecting when is wise not to trade is easier:
I.E. don’t position ahead of a News Release or during a “breaking news”…
A likely objection: 64% predictability looks “too low”.
Consider it includes the broker spread and, believe us or not, is not that low.
Our Research took 12 years to go from 44% to 56% and 5 more years to rise above 62%
Corollary: anyone asserting to be owning an algo whose reliability is higher than 64%…
must prove: he did same research effort , did that consistently and without overtarining.
Example: think of an algo winning an algo competition whose accuracy is 90%.
Our Expertise comes in place. The top 5% of algo competition winners is so overtrained
on that market moment that would make an absurd disaster if used in real market
in a different market condition. Usually the top 5% winner algos,in the long term,
either becomes a decent random signal generators or even contrarians,
Unless someone updates it… but in that case, isn’t same algo at all.
Classic example of Overtraining:
going long on S&P500 while US Government QE was in place,
which boosted (broke) the index to climb only since 2008 ,
just 5% of the index component kept the index rise constantly.
Then such signal is not machine learning at all, is a pure loop:
believing policy won’t change means and face a performance disaster in case it does.
Conclusion:
We expect an Asset Managers performance disaster in the close future:
because all machines since 2008 were trained to expect S&P500 to keep rising.
Which is why We keep our algos constantly under surveillance by the Confusion Matrix instead.