Launch HN: Senpai.gg (YC S21) – Personal gaming coach for PC gamers
It takes a lot of time and effort to become good at popular games. Even after learning a game, you need to adapt your gameplay to meta changes (game updates). We’ve developed a desktop application to accelerate your learning curve, as well as easily adjusting your gameplay as games get updated. We provide personalized recommendations before the game to set up strategies, give instant notifications and feedback while playing the game, and detailed post-game analysis to improve your gameplay.
We're friends from college and played many games together during our dorm life in the early 2010s, especially Age of Empires II (we're old-fashioned). We were very good at 2 vs 2 matches and were considered the best in our dorm. A few years later, we decided to play AoE II again and got destroyed in the first 3 matches. We googled the ways to improve our gameplay and found two ways in addition to orthodox methods such as reading guides. The first was to watch hours of streams on YouTube or Twitch. Considering we had demanding jobs by then, it was impossible to watch long videos. The second was to pay for professional feedback from “human” gaming consultants.
Then, we came up with the idea that we can mimic human gaming consultants and create a scalable system. We could use statistics to provide fundamental level suggestions to improve the gameplay. This is how we formed the basic idea behind SenpAI.GG. Since we know AoE II has a limited number of players, we decided to start with League of Legends and developed a prototype in 2018.
Our Desktop application is now available (beta) for League of Legends, VALORANT and TFT. League of Legends players need to pick the champions, runes and items just before they start a match and play. We provide champions, runes and items suggestions to improve their winning probability. Our recommendations can be imported to the game with a single click. We provide some tags pinpointing the game characteristics of opponents and teammates. For example, we have an "Early Ganker" tag for the gamers who tend to gang in the early game. After the match, gamers can take a look at their strengths and weaknesses. For instance, we provide a tag of "Early Game Gold Loss" for a gamer whose gold/minutes ratio is significantly less than the lane opponent in early game). We provide these tags based on only publicly accessible data on the official game publisher API. For example, any gamer can search for the opponents’ game data and conduct a similar analysis.
For VALORANT, we have a Voice Assistant. For example, we start a counter when a spike (Spike is a type of "bomb" with 45 second detonation time.) is planted and provide verbal notifications (10 seconds, 7 seconds, etc..). For healer agents, we notify the gamers if any of the teammate's HP is low. This information is already available on the screen (health bars) but we provide an extra verbal notification. During the match, we don’t provide any information about the opponent.
We follow the guideline of the game publishers and do not share any information or suggestion that is not available on the screen of the player. We consider SenpAI as analogous to a friend sitting next to the gamers, making recommendations based on the data they see. In the case of sports, the best analogy would be an ‘analyst’ or a ‘personal performance coach' that provides some training and insights about the opponents based on the publicly accessible data. An approximate analogy to, say, chess might be that we tell about the chess clock and provide verbal notifications about time spent while the player is thinking.
We have a Desktop application (built on Electron.js) and a web application. We have two sources of data. We have access to the official API of game deve...
214 comments
[ 2.5 ms ] story [ 236 ms ] threadAs I aged/matured/whatnot I find I have more money and less time. It’s great to have a service where I can close a skills gap time-efficiently for money.
For me it's loss of interest, back log of bought games/gaming pc but can't get into the game anymore oh well.
(Maybe BF when that one finally arrives)
As a former CS player (1.6, CZ, Source, Go) I had these maps committed to memory. Now that I am far removed from school (and the maps got way bigger) this isn't super possible so I find I just get obliterated quickly then just peace out. Not sure if this could help with that as it might border on violating some sort of t&c.
I'll keep a look out for more games! Very cool.
It's surprising to me that there aren't more procedurally-generated environments for multiplayer; there might not be as interesting to spectate (because you don't know where to look or what to expect), but they'd probably be much more exciting to play.
Having random maps doesn't encourage that, but does encourage other things like adaptability and just pure mechanics, since you can't rely on having good positioning and map knowledge of course.
Every week, the developers proceduraly generate new maps using a theme / tileset (e.g. convenience store, chemical warehouse, gameshow studio), then remove ones that don't survive a playtest. A dozen or so maps are introduced each week and there's no way to choose one, making the pool too large for memorization to be useful.
At the start of each round, teams get some time to look at an overhead view, draw a plan on the map, and decide which of their limited resources to use (night vision goggles, breaching charges, barbed wire, etc.). While the layout is random, different areas on the map will have unique features, like a vault that can be opened with explosives or by pressing a button elsewhere, an arcade room with dim lights and cabinets for cover, or a bulletproof cashier's window.
I've been following the project for years and bought a copy when it entered early access last year. It has a friendly community (for now). Big recommendation from me. https://store.steampowered.com/app/753650/Due_Process/
mods and community maps are a different story, of course, but they diverge quite a lot from the main game.
The capabilities of your League app seem to be easily accessible via multiple other apps (porofessor, op.gg, League client itself), and during pick/ban, suggesting champions/runes to use based purely on community stats may actually be detrimental (the player may be much worse at a suggested counterpick, than if they just played their best champion).
I think if the coach client can provide suggestions such as the following, it would provide a lot of utility for minimal effort: 1) reminder to check mini-map 2) after a kill, suggest backing vs. pushing wave to tower vs. roaming 3) itemization suggestion, which has a more objectively optimal choice based on team comp/itemization than champion selection.
[0] https://blitz.gg/
As you mentioned, some of our features are already accessible via some other applications but we aim to provide more comprehensive suggestions.
From personal experience, fast paced strategy and action games just require a ton of practice. I used to read up on the best COD weapons and the best League setup but that never really put me in a place of contention against top players. Regardless of how many videos I watched of the top gamers, I just wasn't good.
Question: I'm wondering what special info your app can provide players that will actually help them get better than just playing the game more. E.g. Knowing a person will "gank" more based on a tag doesn't mean the user will know how to respond to that.
Anywho, good luck on the launch! It's a big feat to even release a product - never made it that far myself haha.
For instance in Age of Empires 2, the best way to improve is to look at a replay of your games, and analyse what went wrong. Stuff like "my economy was idle because I was focusing on the fighting" is much more flagrant then, during the game you often think it was just a few seconds, when in reality it could be a whole minute. I could see a lot of value in having an assistant that analyse your game after the fact and give you hints like: you focus too much on fighting, you have too much resource float, you have a big army but it's doing nothing, your army composition was subpar, you should have anticipated the switch to X, ... You could have it run real time as well, but then you get into the "is it cheating?" debate.
Action games are certainly more skill based as you said, and it's harder to give actionable advice.
I've stood over friends' shoulders and done similar live-coaching. It definitely helps.
You're making this sound easy, but in reality it's very complicated.
How to you measure how much someone is fighting? What if they're playing a very aggressive style? What if the opponent is turtling?
What is "too much" resource float? What time period should it be measured over? What if the player correctly prioritized more important tasks over spending resources?
What is a "big army" and what does "doing nothing" mean? What if they're playing a passive style? What if they're actively roaming the map but not fighting?
Etc.
I have tried to create an analysis tool that does similar things for StarCraft 2, and it's extremely difficult because of how variable and contextual everything is.
It's easy for humans to look at these things and interpret the situation. It's way harder to create rules that accurately reflect human interpretation.
I don't think it's hopeless though, the AI already does something similar. As a very naive implementation, you could for instance run your AI logic on the current state of the game and compare your move to the expected AI move. A bit like with a chess engine.
Map knowledge is a key in shooters along with aim and movement.
Btw, thanks for your support, we appreciated it!
noticed dota either made the game more technically friendly or provided more info in-game (either with a battle pass HUD or just more info in the overlay)
cool that yall are making this for every game
https://dotesports.com/league-of-legends/news/curse-voice-ri...
I think you're falling into the classic trap of focusing on something scalable rather than something that's effective. I'm sure that your system provides some value, but I think even a mediocre human coach would be 10x more effective. What you're providing is incremental improvements based on statistics, but often in order to progress to the next level of competition players need to reevaluate their gameplay on a fundamental level (posture, mental fortitude, interaction with other players, mental model of the game, etc).
I think getting into that human element also opens alot of interesting opportunities to become more entrenched in the competitive gaming scenes which opens up other revenue streams.
What I've seen from paid coaches is that they comment on the game as it goes on, for example telling players to attack the enemy more in lane, and I don't see how this would be replicable without access to LoL's private apis
but now I guess that's against TOS of most games, right? i.e. a voice that tells you what to do / how to do better next round. Now that would be something useful to learn faster!
> GET BETTER AT THE GAMES WITH SENPAI.GG GAMING ASSISTANT SenpAI.GG is the best gaming assistant that will help you get better at the games you play. With the help of Artificial Intelligence, SenpAI.GG offers you the best guidance to help you carry your ranks to the top!
I would also say that what it's doing with Valorant is cheating. It's providing additional cues and advantages automatically, that other players have to keep tabs on manually. The other players have to risk glancing at the timer(s) and temporarily divert attention from the other parts of the screen, while the "AI" user can rely on auditory feedback that is not present in the game for everyone (and maybe was a conscious decision from the game developers).
The whole point of competition is to evaluate player vs player performance, not player+personal assistant vs player performance. If you allow uneven things like that, then it becomes an arms race and also becomes boring from both a viewer and participant standpoint. Part of the challenge of games like Valorant, League of Legends, DotA2, etc, is to learn how to manage the sheer amount of information while also engaging in fights/gameplay.
An actual coach "AI" concept could be to provide intelligent information -inbetween rounds-, not during gameplay itself.
https://www.pcgamer.com/south-korea-makes-cheating-in-online...
https://kotaku.com/forget-banning-online-gamers-are-in-legal...
https://gizmodo.com/chinese-police-bust-alleged-76-million-v...
No, it's cheating. It provides an unfair advantage by replacing a skill with a technology. There may be good intentions, but if so they are founded on a total misunderstanding of competition.
I'd be really interested in a product that does more of what a human coach would do.
Even with DBM - there used to be an addon that told you exactly where to stand to avoid different bosses attacks. That got banned pretty quickly because it over trivialized the mechanics.
If I were an investor in this company, I'd be asking the founders what their plan is when they get banned by the major anti-cheat engines.
Even if there isn't decisive legal consequences they can carpet ban whoever they want whenever they want. Which is likely the reputation which causes the landing page of this service to be plastered with "don't worry you won't get banned" all over the place.
It seems very fraught and risky to use for gamers who desire that their accounts don't get banned.
Its not a service I would ever want to use because I actually enjoy games and I find coaching it to be too much to care about, but I think "audio clues to practice" doesn't meet the bar of cheating lol.
The proper inject time (for instance) is just X*N seconds since game start, the second best time is right after that.
> For example, we have an "Early Ganker" tag for the gamers who tend to gang in the early game. [...] We provide these tags based on only publicly accessible data on the official game publisher API. For example, any gamer can search for the opponents’ game data and conduct a similar analysis.
also seems questionable to me. Is it common/practical to do this kind of thing in these games? It seems to me that, while it might be technically possible to get the info, it should be impractical for a human to actually do the analysis, in the timespan allowed by a pregame lobby, right?
I guess it could also depend on how much preprocessing they do... hypothetically we could imagine a game where every match is recorded and saved to a public user account, and a "coaching" AI system that extracts users tags by watching these offline, and then just queries them at match time. This analysis would pretty clearly be information that players wouldn't normally be privy to, despite being available through legitimate means.
If that's not obviously cheating, make it time dependent, make it player dependent, etc.
If that is obviously cheating, what about producing those maps and not providing an in-game overlay?
If a professional team uses data analysis and then teaches their players something like "at 30s, on this map, this player rotates from here to here in 70% of games", that's just good training. If they have this data being live-streamed to their second monitor, it's cheating.
https://www.opendota.com/matches/5501229759/laning
Maybe its just because I play on Sydney servers, but I really find this a bit overblown personally -- and I've ranked up from bronze 2 to plat 1 over six months of playing. I think its honestly just that some people have "on" and "off" games, at least from the hundreds and hundreds of ranked matches I've played.
I feel like gaming stats should have way better privacy options in general.
Data from games is such a good way for people to get into programming and data analysis (Speaking from personal experience).
It's easily accessible, real and has a decent level of complexity. Would be a shame for it to go away because people want to hide their data.
Realistically speaking this kind of stuff doesn’t really matter at a lower level anyways- everyone’s bad and makes dumb mistakes. And it doesn’t help to know about data if you don’t know how to capitalize on it.
I can’t recommend getting into MOBAs/ARTSs solo, but if you have 2-3 friends they’re quite a blast.
In my opinion for non-professional play (casual and ranked), these kinds of tools should be outright banned. The game goes from “how can I improve myself and play better” to “let’s expend the minimum effort possible solely to win”. You’re not playing to your strengths but just to the opponent’s weaknesses from the get-go. That’s just incredibly boring. Just because one can have a shortcut to a win doesn’t make for better or more interesting gameplay.
Regarding professional play, the teams have the resources to do this given the limited player pool and they’re also at the pro level anyways.
There is a lot of summary stats available, but the most useful one from personal experience is the win ratio. If the data tells you that one player on the other team wins 70 % of the games, then that's usually a clear indicator that the ranking of that player is most likely lower than it should be.
It might just be a better or worse player overall, so adapting to it might not be possible, except for knowing who to put pressure on.
However, "Coaching is available in private lobbies, casual matchmaking, and Co-op bot matchmaking, but not Ranked matchmaking, Team matchmaking, or Tournament lobbies."
The relevant excerpt: "Exposing information that’s intentionally obfuscated (cooldowns or timers)"
As for all of the other match data mining, that's par for the course in competitive games these days.
Doesn't this just benefit people who are good at counting ?
There is a decent sized list of all the timings you should keep track of with in a game, and they change from game to game. If you don't do it you're basically lost.
Doesnt having a good memory give advantage to chess players?
Or having good reflexes benefit shooter player?