It's a circular reasoning that you're giving.
Firstly, understand what seeding really means. More precisely, the idea behind seeding and different variations of Seeding like 'Bucholtz' etc. In chess, there are different ELO's for different players. Do understand that Chess is an individualistic game. A player with 2850 ELO (World No.1 usually) is likely to beat a 2787 ELO.
In CSGO, since it's a 'team' game, it is fundamentally flawed to assign an ELO to the team since that concept is individualistic and since players also do change their teams, ELO concept is thrown out of the window.
In chess, 2800 ELO > 2790 ELO > 2750 ELO. Now, when 2800 ELO vs 2750 ELO happens, 2800 EOL most likely would win (with greater odds than the player with 2790 ELO over 2750 ELO) since the performance depends ONLY on the individual.
But a team has 5 players (+coach) and there are 5 fundamentally different identities. So, the performance on a given day is not reflective of the sum of 5 players since CSGO is a team game.
Moreover, since a player (let's say an average grandmaster i.e. >2500 ELO) in chess plays so many official matches, ELO rating can converge to the 'true' potential of the player in few matches. But in CSGO, since every 6-7 months you have teams with new rosters/changes, an RMR with (let's say) CPH Flames who went 3-0 and played only 4 maps against teams that themselves didn't have a stable ELO rating due to roster changes, the calculated ELO is not accurate AT ALL since we have so few data points.
And since team with newer rosters continuously improve in their initial growth phase, the ELO assigned using only 3-5 maps is not even accurate and I would say is pointless even after 2-3 weeks of the last official.
Now, on top of that you have ELO ratings of teams from different regions (Europe, Americas and Asia) where you can't even compare the ELO across the regions. For example- Complexity with ELO of (let's say 2000) is nowhere close to NiP with ELO of 2000 since they obtained their ELO's from different regions and in reality, if they meet, their current ELO is a poor predictor of the result of their actual match.
I think now you should understand why assigning ELO and using any seeding using these ELO rating system in CSGO is completely pointless since it's a different game with different dynamics of the team altogether.
Also, if you still aren't convinced, then in chess, when we have a Swiss format, then all the players in tournament plays the same number of rounds. After the last round is over, the team with the most no. of points at the end of last round is declared the winner.
In CSGO (rmr and major), Valve is using the 'reduced' Swiss format which as clearly explained above, doesn't properly predict the ELO. On top of that, Bo1's and Bo3's results are also considered the same which implies that the team which wins the Bo1 should also win the Bo3 against the same opponent. So, the fundamental premise is false which violates the spirit of Swiss system (together with ELO ratings) altogether.
That's why your arguement is circular.