HN user
So to summarize, the opportunity of pattern trading is self-fulfilling, which also means if any become unfashionable then the signal will cease to exist. From that fact alone we can conclude there is no basis in the economic reality of the business or its public stock. Greater fool theory, essentially.
Just wanted to say thanks for your transparency - very cool to see you here acknowledging the pain points head on and sharing your plan to address them.
What's up, Rohan! Nice to see you here.
Google's cache: https://webcache.googleusercontent.com/search?q=cache:iAFjnU...
You can't guess Y twice...
People hate change, and this one is very incremental. Without the press release I doubt I would've noticed.
I have had the same experience picking up cycling during the pandemic - work is the last thing on my mind while pushing hard up a steep hill. Cycling comes with occasional injuries too, but is certainly less impactful than other forms of exercise. It's always encouraging to see (and often, be passed by) a much older cyclist, shows that this is a hobby that can be enjoyed for a large part of your life!
I tried yours out, nice work yourself! Seems we took a similar approach in recalculating the letter distributions based on remaining words - both our algos solved it in 4 turns today.
If I may make two small suggestions as a user, I noticed you have a dictionary with nearly 13k words which often results in invalid suggestions like 'clery' and 'meryl'. In testing I found the Scrabble dictionary to be much more likely to yield valid Wordle words (found here: https://github.com/redbo/scrabble), though the official Wordle answers tend to be an even smaller set of ~2,500 common words.
Second, though the implementation is very clean in code (much more concise than mine!), I found the use of the green/gray/yellow methods to be a bit cumbersome when adding constraints. You could wrap these three in a method like guess(word, reply) where your response encodes the feedback as something like [g]=green, [b]=black, [y]=yellow:
Given: [('arose', 27122), ('aeros', 27122), ('seria', 27095), ('riesa', 27095)]
>> w.guess('arose', 'bybby')
vs.
>> w.gray('aos') >> w.yellow('r', 2) >> w.yellow('e', 5)
You could even have the guess method trigger a new round of suggestions since the response implies that we've advanced a turn.
It's a giant array, tough to miss
This is a really cool approach, definitely did not think of trying this! If you'd prefer to play without the crowdsourced data, I spent a couple hours on the following dictionary search algo yesterday which can typically solve puzzles in 3-4 guesses: https://github.com/rgkimball/wordlebot
Nice work! I spent some time playing with this today and used it to build a solver algorithm. It's a greedy search through the possible combinations that prioritizes words based on letter frequency and uses the hints provided to whittle down the search space.
Though not enforced by the Hyperwordle clone, the script plays the game on "hard" mode by default where all letter hints must be included in subsequent guesses. Ironically, this constraint made the algorithm more efficient, easy mode tends to take longer and fail more frequently!
Anyone have a better algo yet? https://github.com/rgkimball/wordlebot
Perhaps "less" is the operative word here.
Is shaming the author really the most productive attitude?
Exactly right - BlackRock and other large indexers have no legal obligation to follow the recommendation of an independent adviser (like ISS, Glass Lewis) in a proxy contest, even for a passive fund. In most cases they passively anticipate the index provider's (S&P, Russell, MSCI, etc.) moves for corporate actions, but management elections for the board and executives are much more subjective as they don't influence the position itself (note that the manager can sometimes also be the index provider, like BlackRock = iShares). The smaller index managers almost exclusively follow the proxy adviser's recommendation because they don't have the resources to analyze board decisions for ~8,000 different companies.
How does this compare to other interpreters?
This is an awfully harsh critique for a free browser game made in someone's spare time.
I worked on this game when I was in High School and at the time we were just entering the first alpha release (amazing to see it grace HN!). Most of the team then and still today is comprised of part-time volunteers, students - people who are simply passionate about game development, art, history or some combination thereof. Many of the contributors actually were sourced from Age of Empires/Age of Mythology fans that wanted to build something that paid proper homage to the things they liked and overcame the things they didn't. Though my contributions are quite distant at this point, I can say that as this is our collective passion project they likely never intend to "release" the game because they're simply having too much fun making it. That would imply a finality that we can instead avoid by passing the project to the next generation of ambitious developers and artists as I once did to my successors. The game is actually quite stable now, supported on all three major OS's, and has a small but devoted modding community. Even though it's free, I've felt that the limited budget (exclusively donations) has always made marketing the game its biggest challenge to finding a broader audience. Hopefully landing at the top of HN will help! We just want people to enjoy the game - if you're an RTS fan, give it a shot!
Big shout out to any of the current or past contributors reading this post - you all deserve immense credit for your work over the decades: https://trac.wildfiregames.com/browser/ps/trunk/binaries/dat...
This is a good start, but given the breadth of applications this would hardly give us enough to compare, as the goal of these models isn't to simply recite Wikipedia articles. What about language translation? Content summarization? Code generation? Turing test performance?
Mason Egger's talk on the topic is transformative: https://www.youtube.com/watch?v=4SwdVMKhbn4