I have slow reflexes, no experience, no expensive gears, limited time and somehow, I managed to be good at a game that I barely played because I have one advantage: I’m a tech entrepreneur.

In 1999, my life changed. I played a mod for Half Life called Counter Strike. It was the first time I ever played anything online and it was the first time I spent my time on a pure competitive video game.

Counter Strike, in many ways, defined who I will become. I built my first high tech company around eSports. I also met my co-founders playing competitive gaming. I had the idea of my start up while trying to find a way to watch better players playing. But while kickstarted my career, I stopped playing it. For many, many, many years.

About a month ago, while I was doing business in Japan with my latest start up, I figured out that it was cheaper to stay in a cyber cafe than staying in a hotel. We eventually still booked an AirBnB but the time we spent in the cyber cafe made me realize few things: 1) no game matches the feeling of Counter Strike and I absolutely enjoy playing it 2) my co-founder spent thousands (literally) of hours on this game and he is good and 3) I hate to be bad at something I like and, above all, I want to beat my co-founder.

So the first thing I did, considering all the limitations I had (especially the hardest one: I don’t have a lot of time), was to learn how other people became good at it.

After lurking for a while on Reddit, I found 2 great youtubers: WarOwl and Voo. They are both analytical and gives great explanations about what they do in different ways. The first one does good editing (and I am a sucker for good editing) and the second one is extremely detailed.

I watched a lot of videos and there are few concepts that I immediately learnt.

The first concept is to think of T-side as water and CT-side as dams. The goal of the CT is to prevent the T from flooding the map. And every time a CT isn’t guarding a specific place, there is a leak and the water potentially goes in.

The second concept is to think that a great CS player is a mix of 3 essential skills: mechanics, game sense and communication.

The third concept is that CS is mental-heavy which means that a player’s mentality during the game will impact this one’s performance a lot.

The fourth concept is that during online sessions, in matchmaking, you usually don’t choose your team mate and they can be either excellent or awful.

The fifth concept is that CS is extremely unforgiving.

Knowing all of this, it is time to do some practice and test how these concepts translate in-game.

I played 3–4 games. I was glad I still remember few things from the late 90s that still apply to the newest version of CS. I also realized how bad I was.

A lot of comments on Reddit to “help” noobs are to tell them to “play the game” and somehow, “with experience”, they’ll be good. I really hate when we use “experience” as a synonym of “magic”. And if they aren’t calling it magic but actually think that by playing a lot of games, the player will figure out how to improve, then, it is basically asking a human to do what AlphaGo did without all the processing power a machine has. There is a reason we do machine learning… with machines.

But how should I improve then? No wonder why my co-founder has thousands of hours on his Steam account. The problem is if I were committing myself to do it that way, or I have to quit everything I work on or I’m going to spend the next 10 years playing this game before I reach an acceptable level.

I did what I typically do when I’m blocked. I decoupled the problem.

The reason it takes so much time and effort is because the concept #2: it is extremely hard to improve in 3 things in the same time. They are disparate skills and they need disparate trainings.

Now there is a concept I learnt from playing poker: adjust your play to your rank. Basically, it if you play against someone and overestimate this one. It is possible then to commit on hands that this one doesn’t get so you lose. Same thing is you underestimate your opponent and this one actually reads your hands better than expected, so you lose.

In CS, this concept can be applied in a direct way: a good player calls for a strategy that no one in the team actually gets. The setup is slow and the execution is inaccurate. The team loses.

The concept can also imply a more indirect observation: depends on the level you play, certain skills are more important than others. And no matter what we say, CS is still in its core a shooter. Yes, knowing how to aim and hit the target is crucial.

On a high level, I’d define the skills requirements as: from Silver1 to Gold Nova Master, proper mechanics (aim, flick, track, movement etc.) are needed. From Master Guardian 1 to Legendary Eagle, game sense is added to mechanics. From Legendary Eagle Master to The Global Elite, communication joins the mix.

In other words, because I’m starting and I’m matched by the machine with and against low rank players, if I can have great mechanics, then I should rank up.

And the best part with mechanics training is it is all about grind.

I had a discussion with a friend of mine who almost turned pro in LoL. Due to family pressure and, well, “life”, he declined the offer and went to college. I asked him why don’t we build a training software to help noobs to be good. And his answer opened my eyes: “the softwares exist but noobs are noobs because they want to stay noobs”. Ok, this needs a bit more explanation.

A great part of video games, almost any video game, is about “eye-hand coordination”. In everyday life, this coordination is needed all the time. For instance, when we draw, write or click on an icon with a mouse. Have you ever played video games with, let’s say, “old dad” or somebody who’s never touched a video game before? Usually, what happens is you facepalm because this one can’t even hit the jump button properly in Mario and dies miserably in the first stage. The reason isn’t that this one is dumb. The reason is in everyday life, it is rarely required to have precise and fast eye-hand coordination.

Now, unless you naturally gifted with great reflexes and amazing eye-hand coordination or maybe you have done rally training and played other sports that require this specific ability, you probably suck at it. From my observation, common people are, by a large majority, not fast and precise enough for competitive gaming.

As my friend told me, the softwares already exist. For all the popular competitive games at least. In CS, for example, there are offline maps dedicated to improve specific skills.

But, and here is the problem for many, it is super boring. No kidding. The best way to improve mechanics is to repeat the same action over and over and over and over and over and over again. You build what is called “muscle memory”. It takes more or less time depends on the person but anyone who’s ready to put the time and effort will improve.

And when my friend says “because they want to stay noobs”, in reality, what he wants to say is that a lot of players know what to do but don’t want to invest the time and effort to do it.

And it is understandable. If you only have few hours to play every few days, you probably want to spend time having fun rather not clicking on target moving around a flat orange wall.

People are afraid of grinding.

I’m not.

A big part in why people are so afraid of grinding is the fear of a bad outcome. People don’t want to invest resource in something that turns out to be a bad investment. Of course. But in the same time, people are impatient and greedy. Sam Altman wrote on Twitter that “we overestimate what we can do in one day and underestimate what we can do in 10 years.”

What usually happens is someone has the right ambition and start to do the training. And then, after few hours, and without any way to measure improvement, this one goes online to check the progress in a matchmaking. If there is a victory, this one thinks the improvement was quick and somehow, this player is now good because, probably naturally talented, and the training is amazing. If it is a defeat, this one tilts and thinks this is complete BS and even after so much training there is still no progress so probably this game is rigged or the training is wrong.

Blame the game (or the other players), blame the training program, blame the internet, blame everything but him/herself. Impatience and greed really don’t work well together.

And, bad luck for me, I am super impatient and ultra greedy.

So to deal with it, I forced myself to not play any online game for a month. I will only use offline maps to train.

I determined a framework so I can track my progress. Considering my work and other activities, I gave myself 1 month and I created a daily schedule. I progress better with “short” yet “intensive” training rather than the opposite.

Mechanics can be split into few categories: 1) Aim, simply put: your ability to “click on a head” 2) movement, your ability to go to precise direction while multitasking (like aiming while moving) and 3) placement, finding and staying at positions of advantage.

I also choose to only use 2 weapons: USP and AK47.

The USP is precise in long distance and is a semi-automatic. This means that during the training, especially in aim, if I don’t hit the target, it is probably my ability and not the weapon. Plus, gun round is important in CS and the USP is the default CT weapon. If I can constantly HS with it, it gives a huge advantage to my team.

The AK47 is, by far, the most impactful weapon in CSGO (and arguably in every CS). Its one-tap potential gives a huge advantage to mad aimers. And aim, is part of my mad training.

I start with aim and to improve this part, I use 2 maps: Training Aim CSGO to improve flick shots and fast aim; and Aim Botz to improve tracking.

In Aim CSGO, I use fast aim mode, the size of the target is on “head” (basically, the target is exactly the size of the head of a CSGO character), I choose to start with 25 targets popping up and at a delay of 0.25s. I also realized quickly that depends on the FPS, the target seems to pop up faster or slower. It is slower at 300FPS and faster at 999FPS. I decided to start with 300FPS. And I use the USP.

In Aim Botz, I use an USP to hit 100 bots, running randomly at speed 3 with neither kevlar nor helmet and with HS only. I activate “show impact” so I know where I missed and by how much.

I created an excel table to follow all my records.

And I started.

And I was miserably bad. I mean, my scores were abysmal.

On Aim CSGO, I spent in average 30 minutes per training. In my first training, I did 43 times the exercice. I got an average of 12.20454545 out of 25 tries. 48.82%. In the second training, I got worse, 47.85%.

It was horrible because I couldn’t even focus long enough. After each exercice, hitting 25 targets, I had to take a long pause. It was depressing. I couldn’t even follow the target with my eyes. Every time I hit something, it felt like luck to me.

In Aim Botz, it was a bit “funner” to “play” since I was shooting moving bots instead of a freaking target. But it was equally bad. If not worse. It took me forever to hit 100 HS. I got an average of 6.04 kill per minute and I got… 19.46 shot per kill. In other words: it took me 19.46 shots, in average, to hit an HS… omg. On my second try, I got 4.3 kpm (basically, it took me more than 23 minutes to kill 100 bots…).

I wanted to stop CS already.

But I didn’t.

Because I told myself that 1) I don’t have enough data to really assess my level and maybe I’m not that bad and maybe I was just unlucky and 2) I invested $25 in this game and 3) I cannot lose against my co-founder. Seriously.

My ability to blind myself, my avarice and my ego kept me on track.

After a week, about 8trainings, I got: 48.82%, 47.85%, 53.15%, 56.90%, 57.53%, 58.18%, 58.65%, 63.03%. I was clearly improving. Almost 15% improvement. In a week.

But more than the %hit, I realized something else: I kept my training 30 minutes. I started by doing 43 exercices. Then, I did 51. Then, 54. By the 8th training, I was doing 77 exercices in 30 minutes. In other words, I shot 1925 targets in 30 minutes and hit about 1213 of them.

I wasn’t that tired anymore. I could focus way longer. I didn’t need to pause between every exercice. My eyes could clearly distinguish a target popping and when I hit it, it didn’t feel like luck anymore.

In the other program, I also see improvement. From 4–6 kpm, I got around 9 kpm in less than a week. That is almost being twice better than before.

And of course, full of confidence, because I’m weak and stupid, I broke my training code and I went online to “frag some noobs”.

I chose a “community deathmatch” server and after few minutes, I felt great. I was hitting every HS. I was owning the game. I was sooooo…. bad.

In my exhilarating demonstration of my ego, I called my co-founder to join the server. Not only I lost all the one-on-one, I also looked at my score. And I wasn’t nearly as good as what I felt I was.

I disconnected from the server and asked myself if I was doomed to fail.

Then, I remembered that 1) I play in 999 FPS but train in 300 FPS 2) I only trained aim and I haven’t done anything with movement and placement.

Back to training.

I kept my aim training routine but changed it a bit. I switched to 999 FPS and I also changed the number of target per exercice to 100 instead of 25. I also decided to be more consistant and I from now, I will do 50 exercices per training instead of 30 minutes. That translates into 5000 targets to hit per training.

To improve my movement, I decided to use a “Deagle” in Aim Botz and keep moving all the time. This is still a precise weapon in mid-distance but it requires to be absolutely still when shooting. This will force me to counter-strafe before each shot.

Finally, to improve my placement, I use a “pre-fire” map. I decided to only focus on the map DE_Cache. Basically, there are bots placed in each corner of the map and force you to 1) place your crosshair correctly so you don’t waste time to “find the HS” 2) place yourself correctly so you only have one-on-one situation and not be exposed to be shot from multiple directions. And I only use AK47, no “nade” at all.

It felt almost like a hard restart. Everything changed from what I knew before. I had to reduce my mouse sensitivity by almost 20%.

I also knew this time that improvement takes time. It is going to be a slow, hard, boring grind. But the outcome will probably be good.

I did it for 2 weeks. Without breaking any code. Just boring offline training. Everyday.

In Aim CSGO, I started at 26.32%. By the 10th training, I was at 36.88%. Before I left for Xmas vacations, around the 13th training, I was at 40.08%. 13.76% improvement in 2 weeks.

On Aim Botz, I started around 10 kpm. By the end of the second week, I was around 14 kpm. I started with around 18 shots per kill and improved to around 7–8 shots per kill.

On “pre-fire”, I couldn’t even “finish” the first level (finding and killing all the bots). I “finished” all the levels after 7 attempts. After the second week, I reduced the time it took me to “finish” each level by more than 30%.

I was improving, I was getting better and it wasn’t a sensation, it was a fact and I had numbers showing it.

And I went on vacations. 10 days.

I initially wanted to keep my training because I was afraid to lose it all if I slack too much. I even downloaded Steam and CSGO on my mac. But even after doing all the tweaks, Mac gaming… well, let’s not write about it.

So after 10 days of no CS at all, I went back to my PC and launched the game. I went through my training programs and I got worse. But not a lot. I went down from 40.08% to 35.68% in Aim CSGO. I was around 12–13 kpm at 8 shots per kill on Aim Botz and I could still complete pre-fire map on DE_Cache around the same amount of time.

And after a week, I realized that I got better. I was hitting almost 50% in Aim CSGO. I did a bit more than 16 kpm at 6.35 shots per kill. A new version of pre-fire came out with more complete training and I was able to complete the new levels faster and faster almost every time I tried. Last time I checked, from main to quad, I was averaging 659 points/3.7s/745hp.

On a non-measurable level, I felt like I didn’t need to clearly see the target before moving my hand and clicking on it. I started to use more and more my peripheral vision. My hand was also moving smoothly. I didn’t shake anymore. I also naturally counter-strafed. Even using a pistol I did it naturally. I was also starting to put the crosshair on potential HS position before I actually saw any head. And I felt calm.

I decided to give CS a try.

I went on a “community deathmatch” server to check my aim. After 40 minutes, I could tell that I was clearly better. It felt natural. It felt… slow. Before, it was I never had the time to even see the enemy before dying. Now, even when I lose my duel, I clearly have time to aim and shoot. Plus, I played on DE_Dust2, which I never trained on so I don’t actually know the right spots and I don’t know the spawns.

Then, I went on a “retake” server. On DE_Cache this time. It was harder and I got yelled on because I was “baiting”. I had to google it before I understood what it meant. Basically, I let my teammate die and then, I use the information to figure out the guy who killed this one so I can kill my enemy more easily.

Using my teammate as a bait.

On a team-building, buddy-buddy, everyone is good and we are all equal point of view, baiting would probably be bad. But considering that my teammate may not be good and that this one may not communicate well and also may not be smart, maybe my best strategy is actually to use this one to lure the enemy and win.

And I won. Many rounds. And I got yelled on. At almost every round. So I muted everyone.

And left.

I was ready for matchmaking.

And I lost 3 games in a row. Something was wrong. I missed an element.

So I watched my demos over and over again, taking notes about what I did wrong.

Here’s the premise: during 2 out of the 3 games, I was the better shooter. Every time I was facing an opponent, I the duel. But in the same time, I had only few rounds where I actually faced one player. I also had multiple rounds where I did nothing. I was alone at my corner, waiting… for my teammates to die one after another.

I also had no idea of any strategy. I didn’t know how to throw any “nade”.

I did some quick math and if everything is correct, my team would have won 2 games out of 3 if I got 1/3 more frag per round. More important: we would have totally dominated the games if I managed to not be the last remaining player. In other words: every time I was in a clutch situation, I lost. So my team lost.

I had some assumptions about why it happened. I wanted to know more. I went on Youtube and checked more tutorials. This time, I watched many videos about “road to the global elite” from different players. Including WarOwl, because I really like good editing.

CS is a 5v5 game and there are many things one player doesn’t control. For instance, I don’t control my other teammates. I also don’t control other players. I cannot see the future. And the recoil, no matter what you do, has some element of imprecision.

Now, the bad way to deal with it is to waste time blaming and changing things we don’t control. The right way is to try to do everything we control perfectly and think of a strategy to minimize the impact of what we don’t control. And yes, this is exactly what we do as start-up founders by the way.

In CS, the bad way often translates into some players wasting their energy and time yelling on their partners or trying to school the others. It doesn’t work. CS is extremely unforgiving and it is skill-based.

At low level, because a lot of issues is related to mechanics and good mechanics only come with grind, telling someone how to shoot won’t help. It would probably do the total opposite: piss off everyone.

Another example is when somebody repeatedly force the other teammates to apply a strategy that is clearly too hard for their rank. As a student pilot, every time the tower gives me more than 3 instructions at a time and I have to repeat it back, I probably miss one out of them 50% of the time. And I just have to repeat it back, not even execute it. So imagine a random person telling 4 other random people to do a 5-step take on A and everyone has to execute in a coordinated way. No way.

I had to find a way to maximize my value and minimize my mistakes while also trying to make other teammates’ mistakes as irrelevant as possible. And after watching so many videos of good players ranking up, the answer was clear.

The best way to rank up quickly, especially at low level, is to mad-carry.

Here’s the theory: if 2 teams are equal in skills, the team that has more players would win in the long run.

If I can manage somehow to put my own team in a man-advantage early in the game, my team will have better chance to win. In other words: what if I could frag very early, almost at the beginning of the round. And as long as I keep my fragging positive, my team should win in the long run.

With this new idea in mind that I started doing more research.

I watched videos from 3kliksphilip to have a better understanding of the game engine. That gave me a good benchmark of what is possible and what isn’t.

Then, I went to train on “pre-fire” map with a whole new mentality: try to be over-aggressive.

I learnt from Thorin about the “positions” in CS. I figured out that being an “entry fragger” would serve my purpose best. So I watched a demos of pros holding that role on DE_Cache.

After few days of research and training, I was ready to give another shot.

Here was the plan: On T-side, I will always take B. It is quite hard, especially arriving at the site, but I found many paths where I’d be only exposed to one side, which gives me a good duel situation. My goal is to frag so as long as I don’t meet anybody, I keep pushing. For instance, if I cleared the site B, I immediately push to Z hoping to catch a CT on rotation.

On CT-side, I’d push hard from A. I learnt 3 grenades per side. And only “solo nades”: they serve me. On CT, I always start with the smoke, then I flash to push in main. If there is no head around, I’d keep pushing to mid. Again, my goal is clear: frag as early as possible and as much as possible.

It worked well. It is easy to understand why. As I guessed right, at low level, players have bad mechanics. So even if they do everything right and managed to somehow see you or double-team you, because they don’t aim, move and place themselves correctly, there still a huge window to kill them before they kill you.

My strategy is to be an over-aggressive entry-fragger. That implies: 1) I force the opposition to react and to adapt to my play, 2) because I have usually better mechanics, I win the gun-fights and 3) I’m almost never in position to clutch.

The last part is important. We got to remember that at this point, I only improved my mechanics. I have no game sense and no communication skill.

Because I have no game sense, no matter how well I shoot, I’m almost always in disadvantage in clutch time. I just don’t know where players are. Worse: I forget how many players are left and when I remember to check the HUD, it is usually too late. Oh, and I forget to check the radar all the time.

On communication… well, my type of play isn’t… always appreciated. So, I cut the voice off. Yes, it is very bad, I know. But I prefer to focus on my play and “click on heads” so it helps my team rather than be yelled on, tilt and do wrong the only thing I know how to do well.

All the reasons to not even try to be in clutch position.

My games go like the following: we usually win the gun round, especially if I get an USP. Then, I save to get a riffle (AK or M4) as soon as possible. I throw the same nades and take almost always the same path. I usually get the first frag. If I get an AK, I usually get more frags. If I’m lucky, I ace the round. If I’m bad, I usually die very quickly. If I die, I write (or, if I haven’t yet muted everybody, tell them on the radio) how many, where and what weapon I saw.

There are also things I do that is probably bad. For instance, if I don’t see anyone in my team who’s carrying the team, I’ll stick to my gun. If I have the choice between picking up a AK47 for me or to pick an AWP for our sniper, I’ll almost always go with option A.

Another horrible things: I don’t care about what my teammates buy. I learnt a rule of thumb from of one Warowl’s videos and I religiously apply it. That’s it.

There are probably many other things I do that are completely wrong. No matter it is on purpose or because of my lack of game sense or game knowledge in general, the truth is: I focus only on what I do well and I created an entire play, for me exclusively, so I mad-carry.

It worked well. It worked even better than expected.

I like to give numbers and share some pieces of data because it helps my reasoning. I also think that if you are reading this and you want to improve yourself, having some numbers could help you to set some benchmarks. But I will not tell you how many games I won in a row nor how fast I ranked.

For few reasons: 1) I don’t understand the ranking system enough so my raking speed may not be exactly related to what I wrote here, 2) there are always elements of luck and other factors I didn’t take in account in someone’s rank so I don’t want to give any false expectations, and finally 3) the rank isn’t a definition of who you are but a mere tool that estimates your average play so it can pair you with other players that hopefully you’ll have fun to play with/against.

Still, I understand the frustration of not having any kind of guidance. Let’s face it: if you were following this training, it could be very, very, very, very, very boring. And I understand that many people need an end goal, or some king of expectations, some king of “potential reward”, before committing to anything repetitive.

If you really need to build some expectations, there is a video from 3kliksphilip explaining the relationship between victories and ranking. For instance, what we see here is that a DMG has an average win of 303. Now, I would assume that if a player is actually better than it’s average teammates and opponents, which should be the trigger to rank up, them this one would probably win more and lose less. So referring to the same video and assuming that a player keeps 50% victory, it implies that a DMG would probably have played 606 games in average. 3kliksphilip did a candle graph representation of the fluctuations that also gives an idea of what can be expected.

Comparing to what I did, I’m fortunate to be on the bottom end of the candle. Especially in the very beginning, at very low level, I managed to rank up pretty fast.

Again, don’t try to hard to compare your ranking to others’. I wrote some good indicators and leading metrics that helped me to improve. As usual, I will be looking for more, better, more precise, with higher correlation and causation.

Ranking is a vanity metric.

Today, I’m at a level where pure mechanics can’t save me anymore. The players I’m facing are at least as good in aim, movement and placement and very often, they are way better. I keep my training, of course, but the marginal increase in my skills is insufficient to win. Also, I almost don’t carry anymore and my type of play doesn’t just annoy my teammates, it is completely gimp.

My next challenges are to improve game sense and communication. The good news is while I was climbing the hills, I learnt some on the way the up. The bad news is it isn’t enough and I haven’t built, yet, a training program that is as efficient as what I did for mechanics. Which is, in many ways, normal.

CS is very demanding. I’m happy that working smart and putting effort gave good rewards. I’m also equally happy that it isn’t enough to master the game. I started to realize how my mental state highly influence my game. I try to keep myself from tilting by playing less and by only playing a pre-defined number of games, no matter the outcome of each, per day. I introduced new “nades” and adapted my strategies to involve teammates. I talk to the radio more.

In CS, there is character level up nor equipment buff, only the you can become better. Sounds good to me. In the end, as an entrepreneur, there are so many things I don’t control but one thing I know how to do is: become better.