Brexit is so confusing it is even confounding the robots.
Machine-driven trading systems in the $5.1 trillion-a-day global currency market are struggling to cope with the blizzard of headlines about Britain's efforts to extricate itself from the European Union.
This is making it more expensive and risky to bet for or against sterling.
Prime Minister Theresa May's failure, after three attempts, to get her divorce deal with Brussels through parliament has sent the UK's planned departure careering off-piste.
It has also raised questions over who is in charge and when, how or even if the UK will actually leave.
As a divided government battles a divided parliament over away forward, the chorus of characters who can now influence events has grown.
This is flummoxing news-reading algorithms, or 'algos', which are designed to parse phrases from recognised speakers before executing a trade.
"The model signals are more quantitative driven and rely on historical data feeds," said Neil Jones, head of hedge fund currency sales at Mizuho in London.
"Brexit headlines have thrown a spanner in their works for the sheer number of characters moving the currency on a daily basis," he said.
News-reading algos are a growing part of a wider revolution on the trading floor of banks and asset managers, where machines have supplanted swathes of human traders, slashing costs and boosting the speed at which deals are done - sometimes down to millionths of a second.
Traditionally designed to process economic data or central bank policy statements, some computer trading models have evolved to allow for split-second analysis of news headlines or Twitter storms before executing a buy or sell order.
The problem for the computers is that Brexit is producing too many headlines for them to process.
Reuters, for instance, has published up to 400 news headlines on Brexit per day in recent weeks, up from around 15 on British politics before it became an issue.
Rival Bloomberg has also pumped up the volume of Brexit content by four times since last autumn, running more than 1,000 headlines some days - such as on March 12 when May's deal was defeated a second time, according to a spokeswoman.
The mechanics of how Brexit may be hammered out have also made it more difficult for the computers.
Obscure British parliamentary procedures are now at the centre of policymaking and people who typically would not feature in a computer trading model are suddenly taking centre stage.
John Bercow, the speaker of the House of Commons, for example, sent sterling skidding last month when he stopped May from bringing forward a vote on her deal.
So sensitive is the currency to developments that even a hand signal can affect the price.
On November 6, Britain's then Brexit Minister Dominic Raab, pushed the pound up simply by giving a "thumbs up" after a cabinet meeting - a visual cue that would outfox machines programmed to analyse words.
Raab's market-moving gesture came after the pound had fallen on a tweet warning of a no-deal Brexit from Jeffrey Donaldson, one of 10 Democratic Unionist Party politicians whose support May needs.
Data is elusive on algorithms' exact share in sterling trade, it likely mirrors broader trends - around 70% of orders in all currencies on the EBS platform, a major trading venue, are submitted via algorithms, the Bank of International Settlements estimated last September.
That compares to a quarter in 2008.
There is no official data on what proportion of trade in foreign exchange is carried out by news-reading algos, but three currency traders at London-based banks estimated it was less than 10%.
Given the complexities of Brexit, that proportion is likely to be even lower for trading sterling right now.
Some hedge funds have opted out of trading sterling altogether because the usual models they rely on do not work in the current climate, according to one FX trader at a major UK investment bank.
Their models are based around economic data and expectations for Bank of England rate changes, but those have become secondary drivers compared with political news, he said.
He declined to be named because he was not authorised to speak publicly about clients.
Some banks are ensuring that trading the pound is not left completely to the machines while other banks are using tiny orders within narrow trading ranges to prevent large losses.
"If it was your job and given the complexity of the Brexit story, do you really want to precode something to automatically infer and put material risk on the back of that," said David Leigh, global head of FX spot and electronic trading at Deutsche Bank.
"Probably not," he stated.
Market makers who provide liquidity by offering to buy and sell currencies on their own account also use algorithms to set bids and offers. But confusion around Brexit has made that more difficult too.
To reduce the risk of getting caught on the wrong side of a headline, market makers are programming algorithms to offer a wider spread between the price they will buy - the bid - and sell - the offer.
But a wider spread makes it more expensive to deal in pounds.
The upshot is less volume as investors stay on the sidelines and wait for the political drama to end.
Daily cash volumes for sterling were around $65 billion in February, around 35% less than $100 billion traded before the Brexit referendum in June 2016, according to CLS, a major settler of foreign exchange trades.
Sterling volatility meanwhile, is at the highest levels in more than two years, having more than tripled from end-February lows, even while volatility elsewhere has declined.
Overall, sterling has fallen 13% against the dollar since the Brexit vote but the market is confused about its future direction.
Buford Scott, managing partner at Stelrox Capital Management, a London-based family office which made money trading sterling during the Brexit referendum, said he was steering clear for now.
"Pound trading is characterised by turmoil and risk aversion resulting in wide ranges and largely directionless markets proving to be generally unprofitable for systematic strategies," Scott said.