When you decide on a game to run, how do you go about it? For example, if it's a game you haven't played before, do you time your first playthroughs, or do you play it casually a few times until you feel comfortable putting on the timer? I'm relatively new to speed running, so I'm just curious to know how others approach the process and maybe get some ideas that could help me in my own process. Cheers and Happy Running!

If it's a game I never played, I finish it casually first. I always try to look after a tutorial, since that usually helps me to learn the game faster. But if there isn't one, watching a run and making my notes helps too.

I do this

1. dry playthrough
2. fast playthrough
3. research strats
4. practice runs, IL runs, sometimes glitch/skip practice
5. full speedrun
6. continue grinding full runs

I play lots of very short web/pico-8 games, so my answer may not apply much to you, but for me, it depends on where I am on the casual/serious spectrum for the game in question. Some games I just do one or two 30-90 minute sessions of runs after beating them once casually, other games I take the time to do one or two 90-120 minute "mapping" sessions where I test out different strats before ever doing a run. And then put varying amounts of hours into grinding them thereafter.

(Sometimes a casual speed game gets more interesting to me over time and develops into a game I put more time into, in which case I usually end up doing mapping sessions some time after the first couple sessions.)

A lot of the games I run don't have other runs available to study before I run them, so I often have to come up with the strats entirely on my own. (At least at first.) When I run a game with other runners before me, I watch their runs before doing any of my own to see if I can find any potential spots to improve on and potential strats they may have overlooked. Basically, any spot where I think, "Why did they do this and not that?" or "Could they do this instead?" is something I go and test out for myself.

Or even just spots that look slow or awkward—those are generally more likely to have some sort of improvement waiting to be found. Again, though, the games I play typically don't have many runners, so they aren't as mapped out as the big popular games. If I were to pick up one of the major Mario games or the like, I'd probably just study the guides and strats, putting a heavy grind into practicing those before doing any runs. (Hard to say for sure, that's just how I think I'd approach them.)

I too play lots of short webgames, and my process is somewhat similar to hahhah's. I almost always try to find my own routes for games before doing and runs, or even watching other people's runs. I've found that this allows me to sometimes find strats or alternate paths through levels which are overlooked by other players. In the game Demon Castle, runners used to take a "shortcut" in the first level through a secret passage in the first room. However, I didn't know about the secret when I was routing the game, and I instead tried to optimize the seemingly slower room. I ended up finding a faster cycle, which actually made taking the shortcut the slowest way to beat the level. I also spend some time glitch hunting before running a game, since glitches sometimes be used to save time. It's also a good idea to glitch hunt because glitches can sometimes result in timeloss, or even softlocks. By figuring out how to perform a glitch, I also figure out how to avoid doing a glitch.

Once I've routed a game, I try to find runs others may have done. I check the comment section of games that I find on itch.io or lexaloffle, and I search on youtube and sr.c for runs. If I can't find any evidence that someone has done a run of the game before, I'll just start doing runs. If I find a run of the game, I'll watch it and take note of anything they did differently. Sometimes the strats are the same as mine, but sometimes other runs have better strategies than I do. I use what I learned to improve my route, and then I finally start runs. The whole routing process usually takes a day or two because I try my best to be thorough before doing runs in order to avoid having to deal with getting a really optimized time, and then finding a new strat that obsoletes it.

When I actually do runs, I start out with a few no-resets to get a feel for the game and to set a baseline time for me to run against. Afterwards, I start doing serious runs and keep going until I get a time I'm satisfied with. Depending on how fun the game is and how hard it is, a satisfactory time can take only a couple hours to get or a couple weeks. Usually I stop running a game after a few days though.

Farmer:

1. Find a game with short categories.
2. Play one or twice (game farmers) or play 3-10 times (every other farmer)
3. Never play the game again

It's beginning at a slow pace. I mostly practice before the entire run. When it comes to game capture, I almost got it. However, I'm having issues getting the audio on the video capture working with both Bandicam and OBS. I might have to start doing some speed runs without audio or with a replacement for the audio found on YouTube.

I have no experience learning a speedrun that hasn't already been routed by someone, and usually I have already seen a run of the game I decide to learn. The vast majority of people look into a game's speedrun after playing it casually and I cannot imagine anything more miserable than to time my first playthrough.

So after having learnt a couple of games myself and having seen a lot more people trying to learn those games, I would describe the process as follows (assuming the game you choose to learn already has been routed and has a semi-active community):

Learn the basics:

• familiarise with the categories (read the rules) or maybe ask on the Discord for a beginner friendly category
• compile resources like guides (the game's srcom guides section)
• if there is no comprehensive guide, ask on Discord whether you should follow along with the WR or whether there is a different run with more reasonable strats
• figure out peripheral stuff like whether you need high FPS or a specific input device (freescroll mouse maybe) or specific control scheme for some speed tech
• if necessary make run notes (maybe some are available on the game's resource/guide section or you can ask another runner for theirs)
• once you have a grasp on the route and basic tech, you can do a first run as a personal benchmark (very much optional)

Actually get good at the game:

• segmented practice
• segmented practice
• segmented practice
• doing runs is an amazing waste of time, compared to segmented practice. Repeating the same trick/fight/section over and over is multiple times more efficient for improving than seeing each section of the game only once an hour
• additionally, during runs you will be afraid of losing the run and play safe. If you play safe, you will not improve. You only improve if you challenge yourself to play better than before.
• getting pbs is fun but does not even compare to actually improving at the game
• when you have sufficiently improved, you can go on a pb grind, but when you get your fat pb, you should go back to practice
• in some games doing ILs is good practice, as a shorter run will be more optimised and require you to play at a higher standard than your full game runs.
• consistency is key. You will have commit a lot of stuff to muscle memory and brain memory, and for that you require a large amount of repetition over a considerable amount of time. Particularly at the very beginning, taking a break can undo quite a bit of the learning process, which makes it frustrating to get back into it. Once you have poured a few hundred hours into a game, your skills atrophy at a much more manageable pace.
• practice is a great opportunity to critically analyse your gameplay and to pay attention to the game's mechanics and AI behaviour or whatever else may be of interested for the game in question. Here you have a chance to find new optimisations or alternative strats that work for you. Of course this becomes more relevant the less optimised and datamined the game is.

A lot of very detailed responses! Thanks guys :)

i dont really ever start to run games before ive played it casually, only after finishing it casually i might feel like i want to run this. when i do learn a run i follow a tutorial/run through the game and then start running never resetting thats the best way to memorize the route for me

I can give you some general considerations and tips on how speedrunners usually approach the process of choosing and running a game. Choose a game that you like. It is important to choose a game that you really enjoy playing. Speedrunning requires a significant investment of time, so it is important that the game is fun and enjoyable for you. Familiarize yourself with the game.If you have not played the game before, it is recommended to spend some time on the regular game to familiarize yourself with its mechanics, levels and possible glitches or omissions that can be useful for speedrunning. Watch speedrunning videos.To better understand the technique and strategy of speedrunning in your chosen game, it's helpful to watch videos with experienced runners. Analyze their routes, running techniques and optimization to learn from their experience. I wonder if there are speedruns on casinos? hahaha, on this site I would speedrun https://bookofragame.com/en/ .

Watch the fastest run to see how optimized/difficult it looks. Also look for time saves that might be possible etc.

The guy with the joke answer got deleted LOL