Retake Strategy with Stats and Sun Tzu
I've been talking about a concept with my teams for a while that I've always found a bit hard to explain - recently I've been calling it "N-1" retakes, which I believe to be the most effective way to retake a site on any map in Valorant. I was struggling for a while though to figure out why it works as well as it does. Then I heard a quote from *The Art of War *that explained it perfectly. More on that later.
So what the fuck is an N-1 retake?** **(please I am also looking for better names for this concept, open to suggestions)
Most of you will already know what I mean when I say "retake" but for the avoidance of doubt, I'm defining a retake as a situation where the Defenders stabilise after Attackers plant the site, before attempting to take the space back so they can defuse. The reason this stabilisation part is important is to exclude situations like floods, or fast plants without attackers actually taking site space.
How to calculate N for any given bomb site - this one is Haven A
N is the number of entrances or exits onto a bomb site, pretty much any way that players can get in or out. We'll use the example of Haven A quite a lot because it's a great check of fundamental understanding for Pro teams - everyone knows A is hard to retake, but only good teams understand how they need to deal with it when it happens.
A site has 4 entrances, so N is 4. An N-1 retake means you should optimally look to retake from 3 of those 4 entrances - the issue is that only 2 of the 4 are usually in Defenders' control (CT and Heaven). This is exactly why retaking A is so hard, and why generally teams look to avoid it. However, good teams understand this and work around it by starting a *flank *as early as they possibly can, so they can open up entrances 3 and 4. You'll notice this especially with teams like PRX or G2 currently, and with FNATIC in the past.
I've had and talked about this theory for a while now, and my thinking was simply that it strikes a good balance between giving your opponents lots of things to worry about without spreading your own resources too thinly - but it was always just a theory based on some logic and pattern recognition. Good players also just... know this, instinctually, even if they don't know exactly why.
But now I can prove it!
I recently tweeted about a tool I've been building that allows me to search through every round of Professional play (that we have data for) using player positions on the map.
It's super useful for finding rounds playing specific strategies, but another use is finding the highest success solutions in different situations.
For example, that Haven A retake is statistically the second hardest in Valorant, with a 26.8% winrate. If I adjust to only retake attempts with 3+ defenders and attackers alive (my definition of an "actual" retake), that number goes up to 32.4%, still very unfavourable for Defenders. Apply N-1 theory, and that goes up to 52.2%. That's 152 out of 291 retakes successful.
The highest winrate of these is retaking through Heaven, CT, and Short (64/117, 54.7%), while the most common is Heaven, CT, and Long (84/160, 52.5%).
The round search in action - each solid blue zone contains 1 or more defenders, while the dotted ones contain 0. Limited to post plants with 3+ players on each side
Compare these to the most common retake pattern overall, the standard CT and Heaven retake, and we're down to 35.3% across 600+ retakes. This pattern has held true on every site I've applied it to so far, and it passes the eye test when watching matches too.
Another great example is Lotus B site - a 39% retake win% is already the highest in the game, but going from a CT and Heaven retake, to adding in a B Main flank adds 15% to your chance of winning the round.
Some of you may be wondering "why not just retake through all of the lanes? why do I have to count this N bullshit? where does Sun Tzu come into this??"
Valorant is not War. Let me make that very clear. I am not advocating for use of terms like Schwerpunkt and Encirclement (sorry Anderzz) - but some concepts do apply. In the famous book Art of War, Sun Tzu said:
"When you surround an army, leave an outlet free. Do not press a desperate foe too hard."
"This does not mean that the enemy is to be allowed to escape. The object, as Tu Mu puts it, is "to make him believe that there is a road to safety, and thus prevent his fighting with the courage of despair." Tu Mu adds pleasantly: "After that, you may crush him.'"
For some reason, this works in Valorant too! In my research, trying to go through every entrance to the site left players isolated and often get picked off early, to the point that I had to manually check over rounds to count them correctly. In our Haven A example, teams attempting to retake through all 4 directions succeeded only 15% of the time.
So what does all of this yap mean for playing the game?
In all of my research, N-1 retakes were highest winrate on EVERY site with the exception of Split B. The solution for players is easy, and it's something good Coaches (and jtay) have been teaching for years anyway - if you know you're going to be playing for a retake, start a flank. Don't get greedy and spread yourselves too thin when you get there.
If you really want to be in the top 1% of Valorant players, you have to learn to count.
