What's New On These Pages: January 2026

I have some future plans for these pages, of course!

• Saturday 31st January
  Finished off the Set for Life XML parsing, but I decided to hard-code the top two prize tiers as £3.6m and £120,000 rather than bothering with the JSON file. I haven't added the JSON download links to the data table below because they include the draw number in the URL. The format is https://api-dfe.national-lottery.co.uk/draw-game/results/gamenum/drawnum where gamenum is 1, 2, 3, 4, 5 or 33 (matching the draw types in the table below).

  I've now also finished parsing the Euro Millions and Euro Hotpicks XML, which wasn't too bad. There is a separate script run when there are multiple prizes for the Millionaire raffle - I adjusted the CSV URL in the script, but haven't actually tested it yet against a multi-prize raffle.

• Friday 30th January
  After parsing the XML results for the main Lotto, I realised the XML file stupidly doesn't have the jackpot prize pool figure if there were no main Lotto jackpot winners! I eventually found a JSON feed of results that does have that figure, albeit in "cents" (well, pence for the main Lotto) and, yes, the cents figure ended in 99 and then got rounded up by the Web pages up to the nearest pound - cue similar calculations for my code and that completed the sorted numbers/prizes/winners/machine/ball set parsing for both the main Lotto (and Hotpicks) and Thunderball.

• Wednesday 28th January
  I've now updated my scraping code to use the CSVs listed in yesterday's table below for all 6 games. It was really just a change in the CSV URL for each game - I've assumed the CSV contents are the same as they were for the old official site. We'll see how it copes with the main Lotto and Thunderball CSVs tonight.

  Next up, I'll start working on parsing the XML files for the full results - one interesting thing is that the XML file contains an exact estimate (to the nearest pound) of the next draw's jackpot prize pool. For example, it claims tonight's main Lotto will have a figure of exactly £5,013,960 and it also provides a short-form version of £5.0m which would be more sensible to display I suspect! I might see if I can use this estimate rather than the one I currently do using some calculations on based on previous jackpot sizes for the same day's recent draws.

• Tuesday 27th January
  The new official lottery site is too difficult to parse "old school" (grabbing the HTML and parsing it) because of its extensive use of JavaScript and I didn't fancy messing around with Selenium or Playwright to get a JavaScript-rendered page.

  Lukcily, XML for the latest full draw info is available, though I've yet to find a way to get older draws. Obviously the CSV downloads are still available, albeit at new URLs which include a game ID (mostly logical except Euro Millions gets ID 33 way out on its own for some reason).

  Here's a summary table of where to get the info if you're after parsing it yourself though, as I said, you'll have to grab the XML for every latest draw before the next draw occurs (because it will only return the latest draw):

  Draw type	CSV (180 days)	XML (latest only)
  Main Lotto	Download	Download
  Lotto Hotpicks	Download	Download
  Set For Life	Download	Download
  Thunderball	Download	Download
  Euro Hotpicks	Download	Download
  Euro Millions	Download	Download

  It should be noted that you don't really need the Lotto Hotpicks or Euro Hotpicks CSVs because they just contain a subset of the non-Hotpicks versions (e.g. no non-main balls because they are ignored in Hotpicks). You will need the Hotpicks XML files of course since they have prize/winners info that the CSVs don't.

• Monday 26th January
  The new design of the official lottery site launched and predictably the links for all the pages/downloads I use to scrape the results from have changed (probably along with the HTML surrounding the results too). The CSV download links are particularly annoying - they seem to be JavaScript driven and not a simple URL either. Plenty of code revamping will be needed on my part - obviously tonight I'll be looking at Set For Life first because that's the first draw coming up. More details on my progress will be forthcoming...

• Monday 5th January
  Spent a bit of time working out why so many requests were being made to the lottery CGI, causing high load on the server and actually running out of Apache processes to serve the site (causing delays or timeouts).

  It turns out that despite me putting /cgi-bin/ in robots.txt as a "do not load" URL for any spider (I don't want spiders to ever run the only non-static part of the site), there's plenty of spiders who egregiously ignore robots.txt, fake a browser user agent string and load /cgi-bin/lottery anyway. They then get "stuck" in the never-ending link maze that is the Number Palace page and hence the massive number of requests from these "illegal" spiders.

  The fix was straightforward - none of these dodgy spiders provide a legitmate referrer URL (it's blank), so if there's a request to /cgi-bin/lottery?Action=Palace without a referrer URL string, then the request is blocked. I did this initially via Apache on the back-end server, but later on also implemented it as a security rule in Cloudflare because the bad robots didn't back off trying to load /cgi-bin/lottery?Action=Palace despite them all getitng 403 forbidden code returned.

Previous month: December 2025

© Richard K. Lloyd & Connect Internet Solutions Limited  2026