<< Back to Warzone Classic Forum   Search

Posts 1 - 4 of 4   
Your favorite CW template (SEAD) analysis: 5/10/2022 14:07:40

FiveSmith 
Level 48
Report
I love Small Earth 1v1 Auto Dist (SEAD) template. It offers an unique combination of strategic depth with simplicity. It is also well-regarded staple of Clan Wars, which, I believe, is the only official clan competitive format at Warzone. (BTW, maybe I should have put this thread in Clans forum instead of Classiс forum)

If you wanna be the very best like no one ever was at SEAD, you gotta search far and wide for various information, which may help you. This is what this analysis is for. As there are some major shifts on the CW meta, the time is right to dig into some insights.

Disclaimer: By the responsible use of Warzone API data was gathered on ~41k games from CW (CW slots from Season 2 up to recently) and ~25k other randomly caught games.

1. How popular is SEAD?

SEAD amounts to ~30% of all CW games.



But... we can't play SEAD as often as we would like to, because sometimes CW slots do not offer SEAD. Actually SEAD is only third most offered template in CW history.



Disclaimer: If a template was offered during CW slot but no games were played, then it will not be accounted for in the stat above. That is because Warzone API provides data only on actually created games. (Sorry, fans of "2v2".)

Recently it changed for good, and at least from Season 11 SEAD is the most offered template.



When we analyze only slots when SEAD was available as a template, then the percentage goes up to above 60%.



There’s no sense in going out of your way to get somebody to like you. SEAD doesn't, and we like it just as it is!

Does the popularity change over time? Nope, SEAD is the staple of Clan Wars



Takeaways:
- SEAD is by far the most popular template
- We want SEAD in every CW slot

2. Is SEAD just a roulette and not so skill based as other templates?

Lets compare win-rates for players, who play both SEAD and non-SEAD games.
There are 748 players in the selection, who played ate least 10 games in each category.



If you leave only games outside CW matchmaking system, then I analyzed 111 players



Even without CW matchmaking, we see, that there is some visual correlation. Correlation coefficient is 0.45, which is significant.
Regression slope 0.47, with R-value 0.44. Roughly speaking SEAD is 50% luck and 50% skill.

Takeaways:
- SEAD may be up to up to 50% luck, but skill is definitely involved
- Success at SEAD does correlate with non-SEAD templates. So teaching/learning SEAD to understand strategic principles is an option.

3. Best and worst starting positions. When should you concede at turn 0?

Analyzed ~15k SEAD games both from CW and non CW which resulted in a Elimination or Surrender (i.e. no booted-wins)

What are you chances of winning if, if you find your starting territories in Europe and Africa? Approximately 45%. See the chart below! Here is your quick assessing tool.



Let's see, how at advantage/disadvantage you are, given we know both your starting regions and your opponents.



That's a more detailed chart showing distinct combinations if territories.
Not all combinations are present, in order to save space. Each combination presented in the chart happened in at least 80 games. If you seek for a combination not present, then I just don't have enough games. But for that read further.



Now we do see the most OP positions, with astonishingly high winrates.

You may notice, that the average winrate of combinations in the chart is higher than 50%. This happens, because there are more available combinations with let's say Asia which is relatively bad for start, and less combinations with Oceania, which is relatively good for start. Thus each distinct combination with Oceania has more occurrences and more chances to get to the chart. And as there is correlation, that part of worlds with less territories are better, then they are more prone to get to the chart.

Let's lower required game count, but will specify territory constraints. Further data was not filtered base on frequency, but at least 8 occurrences are guaranteed for each outcome. Usually the frequency higher.

Truly skilled players should try to win with all their favorites. Best starting territories without SA and Oceania:



Worst starting territories with SA or Oceania



Worst starting territories with SA and Oceania



Takeaways:
- You are guaranteed to not be at a disadvantage if you get both SA and Oceania
- But even if you don't get SA and OC, you can win in a SEAD game. Just face your starting position as a challenge along the way.
- Never concede at turn 0. Yes there are combinations of your and you opponent's territories when you are way crippled, but at start you don't know which territories your opponent have.

4. "I got territories X and Y, what is my best move?"

There are bad ways to win and good ways to lose. What’s interesting and troubling is that it’s not always clear which is which.As there are numerous possibilities here and I am too lazy to present 'em all, I would like to examine just 1 case, which I find interesting.

So, you start your SEAD game and find yourself in Venezuela (SA) + Alberta (NA). Which territories, will you attack. Will you go for Brazil?



Think about it.

If you chose to attack Brazil, you are the majority. Data shows that 2/3 of players go for Brazil on turn 1 and that way they sacrifice all the advantage, that SEAD gods have granted them.

That's the data on conquering Brazil on turn 1:

With decision: 0.25 winrate in 44 games
Without decision: 0.7 winrate in 20 games
Advantage: -0.45

What should you do. Take a look at all the outcomes for your turn 1 options.



Yes, if you focus on conquering NA, you have better chances to win.

I find it tedious and of not very visual pleasing to make a spreadsheet where you could punch in your starting territories and get recommendations on how you should act. But if someone can make that or maybe a web-UI with a real SEAD map and bells and whistles, I would share the preprocessed data for it.

5. TOP-20 popular bad decisions in SEAD

For now I proudly present TOP-20 popular bad decisions on turn 1 of your SEAD game (having 10+ games per outcome and disadvantaging you by more that 10% win rate):



That's all for now.
Hope, that was an interesting read. Do you have any questions?

P.S. Yep, I was pretty bored...
P.P.S. Oh God, it's always been my dream that Duel Lotto would be added to Clan Wars pool!

Edited 5/10/2022 18:16:35
Your favorite CW template (SEAD) analysis: 5/10/2022 14:41:01


alexclusive 
Level 64
Report
I feel guilty that this is my first question, but is there really data that most people attack Brazil in that scenario (??????????!!!!!!!!!!)
Your favorite CW template (SEAD) analysis: 5/10/2022 14:55:19

FiveSmith 
Level 48
Report
I feel guilty that this is my first question, but is there really data that most people attack Brazil in that scenario (??????????!!!!!!!!!!)

Yep, just proof checked 5 most recent games from that position: 3 players went all in on just Brazil and Ontario, 1 diversified yet included Brazil in attacks, and only 1 skipped Brazil. All players are 50+ level.
Your favorite CW template (SEAD) analysis: 5/10/2022 16:52:48


καλλιστηι 
Level 61
Report
- SEAD is by far the most popular template

That is a very bold claim. It is more popular than SE1W, but by ~5%.

Edited 5/10/2022 16:54:40
Posts 1 - 4 of 4