NOTE! These are no longer the newest Brexit Diagrams! The new Series 4 can be found here.

Series 3 also worked! Every diagram had a General Election as the most likely outcome, and that is what happened!

After the success of my two previous series of Brexit diagrams (5 diagrams in series 1 in January, 26 diagrams in series 2 between February and April) I took a little pause for a month after the decision was taken on 10 April 2019 to delay Brexit until 31 October 2019. I even told the New York Times I might stop making these diagrams

But due to popular demand the diagrams are back!

This new series of diagrams tries to plot what will happen on the UK side primarily between May 2019 and the new Brexit deadline in October – up until a possible UK General Election in autumn 2019.

For each diagram 4 documents will be produced:
– the XML file
– a high resolution PNG file
– a scaleable PDF file
– the .ods file I use to calculate the odds

All of these documents will be uploaded to this folder each time – so please check there for the latest versions! Images will also be added to this blog post below, with the newest diagrams at the top.

Due to heavy server load this page might be slow to load – there is a mirror here. Also if you want to use one of these diagrams, feel free – they are CC Licensed for sharing – but please download a file and host it on your site, rather than adding to my bandwidth use. Thanks!

Version numbers will change if the options on a diagram change. If only probabilities change, or minor spelling corrections are made, this will be a .x version.

As ever these diagrams are made based on how MPs have voted in the past, public statements of party leaders and other politicians, and using opinion polls as a means to predict how MPs and Tory activists will behave in the future. All of this needs some educated guessing. However the previous diagrams long predicted a Brexit delay as the most likely outcome, and that was what happened – so maybe there is some use to this!


Version 33.2 – 28.10.2019, 1830
Monday’s effort to get a 2/3 majority under FTPA has failed. Today (Tuesday) Johnson will try a legislative route.


Version 33.1 – 28.10.2019, 1830
Johnson has sent the letter to Tusk confirming the Extension. No Deal 31 October now OFF for definite!


Version 33 – 28.10.2019, 1030
Into the week before the Deadline, it looks like the EU is ready to grant a 3 month extension. And Lib Dems and SNP have come up with a plan to get a General Election – that the Tories might go for when their own FTPA efforts fail.


Version 32.1 – 24.10.2019, 0930
V32 assumed the DUP was against the Queen’s Speech, but after I made it they announced they were in favour. Which then makes it unclear how the Queen’s Speech vote will go – it is currently on a knife-edge, currently going against the government by 1 or 2 votes.


Version 32 – 23.10.2019, 1730
What next? Queen’s Speech then Brussels to decide on an extension (or not)


Version 31.1 – 22.10.2019, 2115
Showing what happened – 2nd Reading approved, Programme Motion rejected


Version 31 – 22.10.2019, 1230
Prior to the vote on the 2nd Reading of the Withdrawal Agreement Bill and the Programme Motion


Version 30.1 – 19.10.2019, 0930
Before “Super Saturday” and the Letwin Amendment.


Version 29 – 17.10.2019, 1600
Johnson somehow has a framework for a Deal, but the DUP still will not back it. I admit my V28 thought this was unlikely (any Deal was only .2), but the overall impact is not that significant – Deal up, election down a bit.


Version 28 – 10.10.2019, 1700
After yesterday’s Varadkar-Johnson meeting, today it seems that negotiations to get a Deal are back on. But managing to comply with the Benn Act is now too tight, so this might be the cover Johnson needs to ask for an extension. All this is covered in the latest version of the diagram.


Version 27.2 – 9.10.2019, 1200
A horribly more complex diagram – to account for the Government having announced it is going to ask the Commons to sit on 19 October. And also drawing on the latest developments in the Vince/Maugham/Cherry case.


Version 26.2 – 3.10.2019, 1915
Now adding the idea that 1 Member State can veto an extension.


Version 26.1 – 3.10.2019, 1630
Now we know there will be a short prorogation of Parliament 9-14 October, and a Queen’s Speech 14 October. This has major implications for the order of the next steps on Brexit. The diagram has also been re-drawn from scratch, and time is more rigorously displayed now. In comparison to V26 this re-adds the Vince/Maugham case.


Version 25 – 1.10.2019, 1730
There is a new idea in town – mandate someone other than Johnson to send the Article 50 extension letter


Version 24.2 – 1.10.2019, 0900
In light of Labour now seemingly wanting a 2nd Referendum rather than a General Election. And cracks in the opposition alliance…


Version 24.1 – 30.9.2019, 1345
Minor update in light of opposition parties agreeing to not table a Vote of No Confidence week commencing 30th September.


Version 24 – 30.9.2019, 1345
Vote of No Confidence first, then extend? Or the other way around? These are the impacts on the different outcomes.


Version 23.1 – 26.9.2019, 1115
After a nasty night in the Commons, questioning Johnson – he is not resigning. And he is not extending. So now chances of a Deal by 31 Oct are closer to 0% than to 1%.


Version 23 – 25.9.2019, 1315
Now we know that Corbyn and Swinson want more guarantees that No Deal cannot happen, and Johnson will ask for an extension, before they commit to a VONC… New routes top left of the diagram. Some odds also adjusted in light of Johnson’s statements.


Version 22.1 – 24.9.2019, 1510
Post supreme court judgment… this one is now simpler! (a few text corrections in comparison to V22)


Version 21 – 19.9.2019, 2130
This version is not just one diagram, but three! Benedict Wesson asked if I could make versions for each Supreme Court outcome – I have done that for whether prorogation is justiciable or not (see diagrams 2 and 3 below), as well as making a version of the standard diagram based on probabilities of how the Supreme Court will rule. Also a thoughtful tweet from Daniel Atkinson made me realise some routes were missing from Version 20 – these are added here.

The impact of the justiciable or not question is rather negligible on the overall outcomes – a General Election is still the overwhelming favourite in each case. BUT the major difference is how to get there. If prorogation is non justiciable, Boris Johnson has a 76% chance of being PM at the 17 October European Council. If prorogation is justiciable, that drops to just 26%!


Version 20 – 19.9.2019, 1200
This one has been built from scratch, the day prior to the expected judgment in the Supreme Court. It aims to simplify everything as much as possible, and also to revert to a more strict chronology of events. This however results in 3 different times when a Vote of No Confidence (VONC) could happen! This version also rolls all possible Brexit deals together into one – it does not try to work out what sort of Deal that Johnson could get.


Version 19 – 13.9.2019, 1830
After the Court of Session judgment, and prior to next week’s case in the Supreme Court – note that this is really hedging my bets on how the case will go.


Forthcoming Decisions V1 – 11.9.2019, 1900
A new mini series on the decisions for the fortnight ahead – trying to assess what happens to the prorogation cases coming to the Supreme Court, and the release (or not) of the Yellowhammer papers


Version 18 – 10.9.2019, 1015
The first post-prorogation version – now an election cannot happen on 15 October. So this one is a little simpler than previous versions!


Version 17 – 8.9.2019, 1400
The final pre-prorogation version, prior to the 9 September day of drama in the Commons!


Version 16 – 6.9.2019, 1015
Labour’s position against giving Johnson the election he wants on 15 October is hardening. The idea the Government might Vote of No Confidence itself has been added – Labour might be more likely to go for this as it pushes the Brexit deadline forward to at least 29 October. General Election this autumn still the overwhelming favourite outcome.


Version 15, redesign – 5.9.2019, 1015
No changes to the outcomes, but some major redesign work after these diagrams were featured on ARD Morgenmagazin. Now the thickness of the arrow corresponds to accumulated probability


Version 15 – 4.9.2019, 2330
This one is in light of the legislation to stop No Deal having passed the Commons, and Johnson’s effort to get a 2/3 majority for an election failed due to lack of Labour backing


Version 14 – 4.9.2019, 1130
Post Johnson being defeated by the rebels in the Commons, and whip being removed from 21 Tories who rebelled. Chances of a GE rising strongly.


Version 13 – 29.8.2019, 1600
This one has some minor changes, adding extra routes for Johnson not acting on what the Commons instructs him to do. This makes the chances of No Deal inch up a bit – to 20%.


Version 12.1 – 28.8.2019, 1610
The Queen has approved the order to prorogue. Diagram amended accordingly. Makes no change to overall outcome.


Version 11 – 28.8.2019, 1000
Now we know opposition parties are going to try a legislative route to stop No Deal Brexit first of all, and only then might Vote of No Confidence routes be attempted. This is reflected in the completely re-drawn Version 11.


Version 10 – 20.8.2019, 1200
OK, so I am back from 3 weeks of holiday, and have tried to get my head around what’s happening now – the Yellowhammer leak, stuttering efforts to get a caretaker government to be formed by Corbyn, but his rather categoric statements that a Vote of No Confidence will happen in September. All of this adds up to the chances of a General Election dropping a bit, and the chances that the UK can agree to nothing other than opposing No Deal increasing quite a lot. For the first time some detail of how at least a first VONC would go is added here.


Version 9 – 26.7.2019, 1430
Johnson managed to see his way through his first week. And we have some more idea what he might try to do now, and how the House of Commons will most likely seek to stop him. This is all reflected in the latest diagram – and this will be the last one until 19th August unless something major happens in the meantime. I need 🏝


Version 8 – 23.7.2019, 1700 BST (note times were CET on previous versions!)
So Boris Johnson has been appointed. I have now added what will happen in the days after his appointment to the overall picture. Also I had previously gone along with the UCL Constitution Unit blog’s conclusions re. Johnson and a majority, while here I have switched to this interpretation from Carl Gardner and others (0.8 probability) – all of that has negligible overall impact, but increases the chance of a General Election in the autumn rather than starting to plan for one now.


Version 7.1 – 11.7.2019, 1130
Now assuming Johnson will win (based on this by Political Betting), and drawing on this about No Confidence scenarios by the UCL Constitution Unit, and with more detail on timing from this by Jonathan Lis. There is a whole new section – to try to work out what happens if a PM Johnson cannot command the confidence of the House of Commons – but that ultimately does not make much difference to the overall probabilities – still at 66% chance of a General Election. The only question is whether that is called sooner or later. Also note that the .ods file uploaded to this folder has a new structure – you can change each node within it now, and see how that changes the overall probabilities.


Version 6 – 27.6.2019, 1130
Johnson is strengthening his position. But what happens were each of them to win?

Version 6 – 25.6.2019, 1730
The final two are known – Johnson and Hunt. And Johnson has started to struggle. All on the latest diagram!

Version 5 – 13.6.2019, 1200
Tory MPs vote in the first round, with Johnson pulling ahead. This change is added to the diagram, but makes little difference to the overall outcome.

Version 4 – 9.6.2019, 1200
This is based on growing parliamentary support for Boris Johnson, notably from Brexit hardliner Steve Baker. It also increases the odds that a hardliner that isn’t Boris would lose a majority. The impact of this on the overall outcome? Just +/- 1% compared to version 3!

Version 3 – 27.5.2019, 1900
Some additional routes proposed by Twitter users @wrexit and @susan04071 and added here.

Version 2.2 – 27.5.2019, 1215
Probabilities adjusted based on reactions to the aftermath of the European Parliament elections in the UK. Paths in the diagram unchanged.

Version 2.1 – 25.5.2019, 1945
After Theresa May’s statement resigning as PM, and building on the betting about the Tory Leadership election and the announced procedure as to how that will work. (Version 2 had an error – the .ods was right, but I had put 26% rather than 23% for No Deal)

Version 1 – 22.5.2019, 1600
Published after Theresa May’s statement on 21.5, outlining her plan to get the Withdrawal Agreement Bill through the Commons (essentially a 4th try to get her Brexit Deal approved). This was produced while rumours of May’s imminent demise is circulating but no news is confirmed.


  1. Sorry, this is more of a technical/software question.

    These diagrams are spectacular. I’ve been wanting to map out decision trees with probabilities for a long time but was always frustrated with how badly the attempt went.

    Are you creating the diagram from scratch, or is there some kind of script or process that helps generate the diagram based on the contents of the OpenOffice spreadsheet?

    Additionally, what is this *called*? Is there a better name than “flowchart” that I could use if I want to read more about diagramming things like this?

  2. Have we not come to a point now that everybody should compromise and accept a second referendum as the least worse solution out of this impasse? Everybody can see how bad the political situation has become, and that an early election will not do the trick either: it will likely break the political parties (and the UK) apart. Is that worth violating the strict interpretation of the principle that you can only vote once on an issue (that was so unclear to the general public to begin with)?

    Who stands up to defend and promote this?

    P.M. or is this chaos Boris’ secret game plan, by provoking everybody, knowing what the end-game will be? As an Etonian game player and the previous mayor of the biggest Bremain city, he might sympathize with the remainers, but working from within, he can manipulate the Tories, without taking the blame for it…

  3. One important thing is missing here. The anti no-deal bill is very likely to be passed, but it’s still hard to imagine Johnson asking for an extension. He may resign, call vote of no confidence or even simply defy the law, but requesting an extension would be the worst possible choice for him, as he would be labelled a traitor by many hard Brexiteers.

  4. Great work (on diagram and prediction), would love to see it continued beyond the GE, even if its hard to guess how it will go, that might become clearer as polls come out, and how does it depend on whether he gets a majority, or minority, or loses and if so to who ?

    • At the moment I cannot predict how a GE will go. Once we get a GE I will then begin to work on that!

      • I totally understand. What I was hoping for was some understanding of what could happen depending on what the result of that election is i.e. what are the most likely outcomes / decision points if BoJo gets a majority, and so on.

  5. Andrew Goldberg

    You have been predicting a GE as the most likely scenario since May (and maybe since before that in earlier diagrams), and now today (2-Sep) it looks that a GE is still the most likely outcome. Excellent job sir!

  6. Thanks for this, even if it’s not necessarily what we want to hear!

  7. That One Girl People Aren't Sure About

    These look like the plumbing in my old apartment building.

  8. Wanted to drop a quick note that the effort you put into this, is noticed and really appreciated.

  9. Brilliant work.

    Just one quibble: since “no deal” is the default, come October 31, could we not crash out regardless of a general election or referendum? So it is possible (I hope not likely!) to have both.

    • Rosalind Stewart

      Patrick, it seems clear from the experiences of both 29 March and 12 April that Parliament would have to pass a No Deal Brexit. This has been confirmed by constitutional experts at the Constitution Unit at UCL and elsewhere. I have also heard experts in Brussels saying that they are already preparing for another extension. So no, crashing out with no deal is not the default.

      And thank you, Jon, as ever, for these insightful diagrams! Personally, I’d put the chances of a second referendum far higher. If I were a Conservative PM, I would reluctantly go for a referendum rather than an early General Election.

  10. Paul O’Connor

    That’s a lot of work, well done.

  11. Stephen Archer

    Excellent Jon

    Keep up the good work. It struck me, in more simple terms, that you can’t be in and out of the Customs Union at the same time. The referendum vote demands OUT the Good Friday Agreement demands IN. Two incompatible outcomes, hence the appalling 585pp document called Chequers, a vain attempt to square the circle.

    • Theodore Welter

      Circling the square is possible if you cheat a little and people cheat all the time. The problem is that Ireland would veto such a cheat because the veto would encourage reunification no matter what happens: If UK revokes it makes the right/DUP/UK nationalists look foolish and weak , if non-Brexit Brexits it silences NI in the EU, Permanent (because Ireland would veto everything) backstop cuts off NI from rest of UK or UK crashes out and annoys NI. There’s a good chance that Boris Johnson goes for circling the square, knowing that Ireland would veto, and use that veto to score free brownie points.

Leave a Comment

Your email address will not be published. Required fields are marked *