mariteaux's Best Times (as of 2019/06/01)
If you haven't used a computer since 1992 or so, let me congratulate you on making it to my site in one piece before explaining the basic premise of Minesweeper: there's a big grid of squares. Click on a mine, and you die. Most squares have numbers on them, letting you know the number of mines adjacent to that square. Click on all the squares without mines as fast as possible without blowing yourself up.
For such a simple game, there's a whole slew of optimizations, strategy, and pattern recognition that goes into it. I blame LGR for making me take a second look, and now I'm a little obsessed. I'm still fairly shit at it, don't get me wrong (I am horribly inefficient in mouse and patterns), but I made this page to post my own best times and explain what isn't easily explained, as best as I understand it. There's a lot of numbers here.
(I use Minesweeper Arbiter, a clone that better keeps track of tons more important stats and is what all the official cool kid Minesweeper players use anyway. Just thought I'd mention that.)
Strategy and Terms
Flags? No Flags?
Minesweeper lets you right-click on any square to mark it as a mine. It's super handy for remembering where the mines are, but the game ends when you've clicked all the squares without mines under them, not when you've flagged all the mines. If you want, you can skip flags altogether, ostensibly saving a bunch of time.
In practice, it's not that simple. NF play requires you to keep track of mines yourself, and if you're used to flagging, you'll probably just outright click the mines a few times thanks to muscle memory. At least, I did. Chording (clicking either both mouse buttons or the wheel) lets you clear the immediate squares around your mouse as long as you've flagged the right number of adjacent squares, and that's a big time-saver too. Naturally, that doesn't work in NF play.
This is why Arbiter (and I assume other clones) keep track of F and NF times separately, because the two require entirely different strategies. If you're flagging, you can chord and click a bunch of squares at once, but you're slowed down by the impulse to flag everything. If you're not flagging, you don't have that, but you have to keep it all in your head and click each square, one-by-one. Both are a pretty fun challenge in their own right.
Counters and Stats
Arbiter keeps track of a ton of stats related to how efficiently you clear a board on a run, and you see this in a separate Counters window as you play. It looks like a bunch of gibberish, I'm sure (and a lot of it still is to me), but here's the most important bits:
- RTime: Board time to the hundredths of a second. Parentheses shows the rounded up time you'll see on the board itself. Est RTime shows what it guesses your time to be if you fail the board, and that usually lets me know how much time I have to comfortably shave off a PB with where I'm at.
- 3BV: Bechtel's Board Benchmark Value, or the number of clicks it takes to clear a particular board. Usually, you can tell how difficult a board is by how many clicks it takes to clear it. (3BV doesn't mean the amount of times you clicked, by the way. Clones and recorders calculate the 3BV after the board's been solved with a counting method I still don't understand.)
- 3BV/s: 3BV divided by the board time. This is a good metric of efficiency; a higher 3BV/s means more of those clicks and less time spent idle.
- ZiNi and H.ZiNi: Another measure of how many clicks it'd take to clear a board. 3BV is long known to favor NF players, so ZiNi (or Greedy ZiNi) and H.ZiNi (or Human ZiNi) are two other algorithms for counting board difficulty, but specifically for flag play.
- Left/Right/Double: The number of left-clicks/flags/chords you did on that particular board. Not terribly sure what the @ nonsense after the click number is.
- IOE: Index of Efficiency. Actual clicks divided by board time. Chording makes it possible to get this over 1, but usually, it'll be under 1.
- Corr and ThrP: Correctness and Throughput, two linked metrics for how many clicks on that board weren't wasted, and the IOE for that board if the Correctness were 100%, respectively.
- Path: How far your mouse moved, in pixels. Moving less is obviously better.
Seriously, if I didn't list it, it's probably not that important. Some metrics, like RQP, are really only counted for funsies, no one actually measures anything with that anymore from what I can gather.
Like I said, I am a horribly inefficient player. I scramble and waste a lot of clicks flagging and chording for little benefit, and there's certainly a lot of time for me to cut off all of my PBs (especially the NF ones).
A big part of efficiency is pattern recognition. If you see a certain set of numbers in the grid that always get solved one way, you don't have to think about what to do with them, thus speeding you up overall.
There's a few patterns in this grid you can immediately use to determine where mines are. Any 1s on corners are automatically mines (hence why it's safest to work from the corners in), and the squares directly adjacent (ie not diagonal, in this case underneath) to the 1s at the bottom of this grid are mines. After you've found those, you've also satisfied various 2s in the grid and can uncover more squares.
Should go without saying, but if your client has question marks (?) enabled, disable them. They have no use and will slow you down.
If I can think of anything else to note, because I've likely forgotten a bunch, I'll write more here soon. For now, here's a bunch of competitive Minesweeper related links.
Links and Further Reading
- The Authoritative Minesweeper, which is pretty much the hub for serious Minesweeper play
- Far more about strategy and patterns than I got into here
- Authoritative Minesweeper's forums, in case you're a math wiz and wanna get into all the detail
- Minesweeper Wiki, which isn't as detailed as it should be for some things and crazy detailed for other things