News - Project Euler

During the summer break last year we began exploring a new system to replace the difficulty metrics, which we felt suffered from two problems: heavy reliance on the fastest solvers to obtain the measure of a problem's difficulty, which meant the metric was skewed by more experienced problem solvers, and its static nature meant that it was unable to adapt to the evolution of programming languages; that is, not all X% difficulty problems are equal.

After arriving at a system which we felt had the capability of properly representing the entire Project Euler community, and running the system against test data, we feel that we have been able to develop a system that is going to provide a far more meaningful measure of problem difficulty. But please note, as the new system tracks encounters between members and the problems they solve to help calibrate the difficulty metric for each problem, our privacy policy has been updated to reflect these changes.

Instead of using difficulty percentages, the problems are now given a difficulty level, similar to your own level of progress. Essentially, the entire problem set is ranked by this new metric and then grouped into blocks of twenty-five to create the difficulty levels from level 0 upwards. The Archives table will colour code the levels from green to a light red to show you which problems are most suitable for you to tackle at your current level. This is not something that will work for everyone, but we hope it helps to guide most members on which problems to tackle and which ones to currently avoid.

It is worth noting that, due to the dynamic nature of the new system, which continues to learn the difficulty of each problem, the difficulty level for some problems may change slightly. The good news is that observing the data during this test period indicates that problems quickly find their place and many are now very well established. The challenge will be problem visibility. That is, those problems "in the middle" that do not receive as much attention as others. Over time as they are tackled they may rise or fall, pushing other more established problems up or down respectively. But we felt that these shifts would be limited and it was a small price to pay for a system that was continuously providing a measure with increasing accuracy.

To assist differentiating those with a stable rating versus those still in a state of less certainty, you will see a small coloured bar below the difficulty level for each problem. Green indicates a high confidence in the difficulty level, orange indicates that it is probably in the right place but may shift slightly, and grey indicates less certainty.

In addition to the new difficulty system, the awards have undergone some changes. Upon review we decided that a primary objective is that each award should be reasonably achievable. So you will notice that a number of awards have been removed, like Gold Medal. In the case of this award, we felt that earning the top spot in any fastest solver table was reward in itself. We know that some of these decisions might frustrate members who have worked hard to earn awards that have now been removed, but we hope you agree that the ultimate purpose of awards is to support and encourage all of our members as they work through the problems.

However, with that potentially disappointing news, there is some really good news. All awards are now permanent. To accommodate this, you will see this has been reflected in the slight changes to criteria for some of the existing awards. For example, Master of Archives now requires you to solve fifty problems in ten grids. Still challenging, but achievable.

There are a few other changes you will notice. The Problems tab on your Progress page now uses a more vibrant coloured heat map to show the levels. We have even provided a switch that lets you turn on/off the entire heat map.

As you can imagine, these many changes will undoubtedly lead to bugs. So if you notice any issues, please report them at: https://projecteuler.chat/viewtopic.php?t=8042

We really hope that you appreciate the considerable time and effort that the team have put into these changes to allow Project Euler to continue providing an enjoyable place for everyone to solve a diverse collection of high quality, interesting problems.

Sorry if the extended downtime today interfered with your plans to solve problems. As some of you know, 2038 marks the end of the Unix timestamp. So with 2038 fast approaching (; a large back end update was carried out today to shift all of the internal time/date DB values from Unix timestamps to DateTime objects. This has taken a number of weeks to modify all of the scripts on my local machine to handle the new DateTime objects throughout the site. It has also required significant changes to the DB tables.

I have done as much testing as possible to ensure that everything is working but until the site undergoes extensive use, it is unlikely that I will have covered all possible interactions. If you notice any issues, I would be very grateful if you took the time to report the issue at the alternative forum: https://projecteuler.chat/viewtopic.php?t=8026

I wish the changes today brought something more exciting, but it was a necessary change. So I would like to thank you for your patience and be assured have already begun thinking about how to handle the end of DateTime objects at the end of year 9999.

The technological world has changed beyond our wildest imaginations in recent years with the widespread access to generative artificial intelligence tools. Whereas this is mostly for the good, a small number of members have possibly been using it inadvertently in a manner that conflicts with the mission of Project Euler. So as to be absolutely clear, we have added some additional guidelines that can be found in our Generative Artificial Intelligence Tools Policy.

For most members, it will be business as usual, but there are a tiny minority of members that might need to review how they have been solving problems when it comes to competing for places in the public tables and any materials posted in the problem threads.

We really hope that you appreciate our efforts and we will update the policy as and when it becomes necessary to ensure that it aims to protect the integrity of everything we do here.

Members have always had the option of making their accounts private, which means they will not appear in any public ranked tables, including the fastest solver tables. However, we have now provided an option which allows members to maintain public status, but opt out of individual fastest solver tables. If you have earned a place in any table, by viewing that table you will see a link at the end of the row containing your username which allows you to opt out of that particular table. Please be aware that opting out is permanent, so you will be asked to confirm your choice before it is completed.

To ensure consistency in relation to privacy options, and to simplify the whole process, a change has been made to how accounts with private status are now handled. This change only impacts the fastest solvers tables. If the account is private, any problems solved during this time will automatically be opted out to reduce the burden of performing this task manually each time.

However, an important change to be aware of is that the choice of switching accounts from public to private now will automatically opt you out of all previously solved problems.

Please note that if your account is currently private, no changes have been made to your previous problems at this stage. However, if you would like to take advantage of this more robust privacy option, you can switch to public and back to private again.

These changes are made clear on your Account page in the section where you can make your account private, where it explains the consequence of switching between public and private. In addition, the privacy policy has been updated to reflect these changes.

We hope that the flexibility of the new opt out feature and the simplification of what it means to have a private account is clearer and more helpful for everyone.

The guide to formatting posts has always asked members to use single dollar for inline TeX content and double dollar for block level, but some alternatives were previously supported. For example, it was possible to use \[ and \] delimiters for block level and you were able to add TeX to the page directly by using \begin{...} \end{...}. However, this put a burden on the JavaScript module that processes the MathJax because it had to scan the entire page to render content. Consequently, support for the alternative delimiters has been removed.

Due to the way that the new BBCode module parses posts, TeX content is now served directly to the MathJax module, so this should make pages load quicker and put less burden on client browsers. In addition, restricting the delimiters to single dollar and double dollar enables the posts to be more stable as content is strictly processed as TeX.

Rather than just turn off the alternatives and leave posts broken, I have done my best to convert existing posts to handle the new required format with some degree of automation. But please be aware that I had to manually oversee more than one-thousand posts over several days, so there may be some of those posts that are not displaying as expected any longer. I would ask members who used the alternative methods to be vigilant and if you get the chance, please check over your own posts to ensure they are viewing correctly.

I realise this might be frustrating for those who have been using the alternative methods, but I hope that you appreciate the attempt to make the forum more robust and easier to use for everyone.