A-Book vs B-Book Execution: How Hybrid Execution Models Work in OTC Derivatives

Financial derivatives are a suction funnel from retail to institutional money (from discretionary to algorithmic money).

The curious world of over-the-counter (OTC) derivatives, as well as financial markets in general, revolves around millions of retail traders with discretionary strategies giving billions of dollars every month to the major algorithmic market institutions without the former even understanding that they are nothing more than an absorption instrument offering their liquidity to the major global financial centres.

Virtually all active retail traders in structured products lose money in the medium term, because to be profitable in the medium and long term in derivatives it is necessary to have a systematic strategy with algorithmic risk management -among other skills- and only professionals are able to do this over a long period of time -thus exploiting market inefficiencies.

And since almost all retail traders lose money in almost all instruments and underlings (OTC derivatives, exchange-traded derivatives, spot markets, blockchain securities...), the world of retail trading is nothing but a feast for liquidity providers, market makers, proprietary firms, brokers, quantitative firms, or any financial intermediary which, being the counterparty to the retail trader, earns all their losses. The golden question is who is the counterparty; it is that algorithmic counterparty that sucks the retail trader's money without the retail trader probably knowing it.

The fact that almost all individual traders lose money is the reason why most of the time these retail traders, when they trade, do not trade with "the real market", but with "artificial" liquidity pools created either by the broker or exchange itself, or by other non-bank liquidity providers with whom the broker has profit-sharing agreements to take a percentage of the trader's losses.

In other words, the entire financial market supply-chain understands very well that retailers are going to lose, and all intermediaries are eager to be the retail trader's counterparty, including exchanges and brokers.

Think about it: Why is the broker going to connect a retail trader to the major aggregators and liquidity providers on the planet if they are likely to lose all their money in less than 90 days and it is better for the broker/exchange itself to earn that money, rather than a quant LP/MM in Hong Kong or London?

Even those brokers in the CFD arena that claim to be ECN (Electronic Communication Network), STP (Straight Through Processing), DMA (Direct Market Access), or NDD (No Dealing Desk), are mostly connecting you to liquidity providers and aggregators with whom they have a partnership contract to profit from your losses.

However, if risk is well managed at the broker's or exchange's dealing desk, these practices are not necessarily unethical, but rather obvious from a business point of view.

At the end of the day, what matters is whether your strategy is algorithmic and winning over the long term so that you are in the A-book. If you trade with a professional and profitable strategy, you will be connected to the highest quality liquidity and not be a simple suction tool for the quant market makers.

The hybrid model of brokers and exchanges offering any financial derivative: A-book vs B-book

Brokers that offer OTC derivatives such as CFDs, exotic options and other structured products, operate with a peculiar model that has been quasi-mimicked by other sectors in financial markets, including the spot share dealing industry (the PFOF model, created by Bernie Madoff) or blockchain (the blockchain market works very much like the OTC derivatives market, especially if we go to the famous perpetual futures contracts or synthetic tokens).

The model is to segregate traders into 2 execution categories with 2 different servers: winning traders (A-book execution) and losing traders (B-book execution).

The famous B-book

Losing traders, a category that comprises the vast majority, are directed towards liquidity pools that are controlled indirectly or directly by the broker or exchange itself.

In the B-book brokers internalise all the trader's trades and connect him to their own proprietary liquidity pools, which means that they profit from his losses as they do not outsource their trades and remain in their own artificial markets as the trader's own counterparty.

As we wrote earlier, in the area of OTC derivatives such as CFDs some of these brokers call themselves STP (Straight Through Processing), ECN (Electronic Communication Network), NDD (No Dealing Desk), or DMA (Direct Market Access), but most of them only use these definitions to hide the fact that they are "hybrid" brokers (A-book vs B-book); meaning that even though they connect you to external liquidity providers and that in itself is true, they still profit from almost all client losses by having a contract with the provider.

Ask them by email to show you which LPs they deal with and see how they respond. Declaring this information is confidential is like a dermatologist telling a patient that the components of a recommended drug are confidential and must be kept in private. The reason they won't tell you where they get their liquidity from is not because they are afraid you will implement an arbitrage strategy between different servers, but because with a little research you might find that these providers are controlled by them, indirectly or indirectly.

The B-book model is the pioneer in the entire financial markets industry and it is what we have been saying so far in the article, that retail traders are nothing more than a tool to further enrich the institutional traders because they are not connected to the real market but to artificial liquidity pools that only try to mimic the real market.

If for any reason a high volume trader generates a very high return with certain trades in very volatile scenarios and the broker has him in the B-book, the trader will literally be making a hole in the broker's treasury, and the broker will do everything possible to prevent the trader from subsequently closing those trades or withdrawing profits.

In these scenarios, the broker or exchange can also act pre-emptively: B-book clients could not be allowed to trade in certain extreme volatility scenarios.

The B-book has the same business model as a casino and the reason why every year certain brokers and exchanges go bankrupt is partly due to not properly managing the A and B-book of their clients.

The A-book

A-book execution connects you to very large liquidity providers, mainly banks. They charge the broker/exchange fees for providing their liquidity -mainly spreads or taker/maker fees plus swaps-, and the broker/exchange adds a mark-up to those costs, the difference being the source of their gross revenue.

In this A-book, the broker/exchange generates revenue through spreads, swaps, maker/taker fees, and/or other optional services, such as guaranteed stop losses to prevent slippage. Nevertheless, in the A-book, the broker does not profit from your losses, but from the fees involved in facilitating your trading through their servers to the banks.

If you are a winning trader, this is where your trades will go.

The C-book

Yes, there are brokers/exchanges with what I like to call a C-book. This is a book where there is no real trading whatsoever. Everything is a demo account and due to the fact they expect most traders to lose all their money, what they do is to collect deposits, wait around90 days until most traders burn their accounts, and they just retain client's deposits as gross revenue. No, there is no client funds segregation in this case.

Even if this practice could be considered illegal, you will be surprised of how many OTC brokers implement it. The C-book is quite common in the prop firm megatrend (not the famous proprietary trading firms from the US engaged in share dealing and futures) but the new generation of CFD prop firms charging traders to use demo accounts collecting evaluation fees without offering any financial service.

PFOF vs. non-PFOF: The B-book vs. A-book model in spot products

In spot share dealing, something similar happens as in the CFD industry with the A-book and B-book models. There, the hybrid model is called PFOF (Payment for Order Flow), where brokerage firms sell their volume to quant market makers such as Citadel Securities, Virtu or Susquehanna. This method was created by Bernie Madoff, the fraudster who created one of the largest Ponzi schemes in history.

Although PFOF execution is not as extreme as B-book model execution in the OTC derivatives space or in the blockchain ecosystem, it is banned in Canada and the UK, but is widely legal in the EU, Australasia, the US, and almost all Emerging Markets.

Some of the brokerage firms that do so include commission-free brokers such as Robinhood, E-trade or Charles Schwab, while some of the spot brokers that do not benefit from this PFOF model are Interactive Brokers, Merrill Edge, Fidelity or, more recently, Public.com.

Which broker or exchange to choose?

If you are executing a long-term winning strategy, choosing a broker or exchange that 100% executes your trades in the A-book and connects you to quality external liquidity is crucial. That way, you avoid being a winning trader in the B-book, which is the biggest fear of all brokers in financial derivatives.

Even if I am not an introducing broker, in the OTC derivatives arena (mainly CFDs) brokers like Global Prime, CMC, IG, Tradeview, or Saxo can guarantee you a quality A-book if you have a winning algorithmic strategy.

In blockchain, it is recommended to trade mainly with DEXs such as Dydx or Uniswap, although we could say that Coinbase or Bitstamp can guarantee A-book execution at the same time.

In spot trading, Interactive is the most professional option and possibly the best retail broker ever created. A big kudos to Mr. Peterffy.

The best solution: implementing an algorithmic momentum strategy in brokers that guarantee you A-book execution

The only long-term methodology to beat the markets is to implement an algorithmic momentum strategy that extracts inefficiencies from the market simply by the existence of volatility, regardless of whether it is bearish or bullish.

These systematic strategies are called algorithmic momentum trading -mostly trend-following and mean-reversion, but break-out too- and while there are many bots in the market that have very volatile risk engineering such as grid, martingale, or arbitrage bots, there are also many strictly engineered trend-following and mean-reversion bots that beat the market and have been implemented for over 50 years in US-based managed futures.

The main drawback I see in the perception of algorithmic trading is that those outside the industry tend to think that you have to develop some sort of celestial mathematical paradigm to extract inefficiencies from the market and approach Tesla's genius when it comes to developing your portfolio of bots.

While it is true that at the institutional level quants managing hundreds of billions of dollars execute quite complex strategies (most of them based on alternative data and high frequency trading in black pools), trend-following and mean-reversion models -momentum trading- outperform market returns, are simple to implement, and have a very mature risk engineering.

Likewise, many industry outsiders may have the belief that HFT (High-Frequency Trading) is some kind of evolution from LFT (Low-Frequency Trading). This is not the case. HFT is widely used in quant LPs and market makers while LFT is more used by the institutional buy-side. If there are trends, momentum low-frequency models can outperform the market.

With decorrelated momentum strategies you can realistically achieve over 30% annual returns with risk/reward ratios of over 1:3 and Sharpes of over 2, unmatched conditions that probably cannot be achieved in any investment offering in all financial markets. (This returns are coming from a retail traders point of view without the sophisticated risk profiles of institutional money managers).

And by using a portfolio of different decorrelated trend-following and mean-reversion bots, firstly, you are guaranteed that brokers and exchanges will send you to the main liquidity providers and not to a B-book, and secondly, you are guaranteed that you will be able to generate profitability in bullish, bearish and sideways periods without your human intervention.

In the not too distant future, anything other than algorithmic trading will increasingly resemble the betting industry rather than financial markets. In fact, at the retail level, that era is probably already here.

The main programming language in algo trading is C++ because it is much faster than others such as Python, although most retail strategies are based on proprietary languages of charting platforms such as Trading Technologies, Pro Real Time, Interactive Brokers (C++ & Python), Amibroker, Tradestation, Rithmic, Ninjatrader, Metaquotes, cTrader (C#) and many others.

At the institutional level, funds such as CTAs obviously do not tend to use tertiary platforms due to the need to protect their code, although at the retail level, it is more appropriate.

Fortunately, more and more discretionary traders are becoming systematic, using bots for their trading, and are able to generate attractive, anti-recession returns, which is the opposite of what most traders do and what most intermediaries want: that you don't understand the functioning of finance and lose all your capital with manual trading.

It won't be long before retail traders "wake up" and products start appearing that allow you to implement a portfolio of decorrelated momentum bots without any trading knowledge.

Francisco F. De Troya

Algorithmic trading & derivatives professional.

Executive Chairman, Blockmas