You seem to be missing particular parts of information regarding war scores.
Let’s go back to basics. The war score is dependent on several factors:
- Top 30 heroes of each participating player.
- Top 5 heroes have extra weighting (assumed to be rainbow).
- Top 5 troops (assumed to be rainbow).
- Win/loss record for each participating player (previously win/loss record of the alliance).
TP is a metric which comprises the various information regarding the heroes’ stats and the troops. People who use TP as a definitive measure tend to overlook other important aspects of a team’s (defensive in this particular scenario) composition. It is possible to defeat a team with 500 TP higher than your own attack team with the correct composition and tiles.
The one part of war score missing is the win/loss modifier. As I mentioned above, a player has a base war score comprised of points 1, 2 and 3 above. If someone had the interest I’m sure this can be calculated and generalised. It’s point 4’s contribution to the war score which remains unknown as far as I can tell.
Basically, take that base war score then it gets modified by some factor to create the new war score after the war. Now it should be easy to see this: an alliance who have lost a lot (whether due to bad form or by design) will be close to their base war score.
Now suppose an alliance with a lower war score has been in good form and won a few. Their base war score is modified so it is greater. Put the two together and there’s a higher probability of these two alliances being matched.
One thing you may see often claimed is if you want to avoid mismatches then you need to be in a full alliance. I don’t think it’s completely true. One factor in this is actually the bench/roster difference between the ‘highest TP’ player and ‘lowest TP’ player in the alliance. It’s really simple mathematics: how many different ways can you add up to 20? Take this example and apply it to a player’s base war score. And then apply it to the entire alliance. Notice the potential variance between top and bottom? Personally I believe that variance is more likely responsible for perceived mismatches in strength.
The war matchmaking process is essentially aimed at a 50/50 win/loss ratio. Once you win a few then expect to lose a few. With a higher war score you’ll either meet similar base war score alliances on similar form, or higher base war score alliances on a losing streak. That’s not to say it isn’t possible to be ahead of the curve, so to speak, but it would likely require more strategising and taking the game far more seriously than some may want.
Note also, the new alliance loophole has been effectively closed by the movement of war history to be a player metric rather than alliance. Each player will carry their own war score to a new alliance, so forming a new alliance will no longer reset the alliance’s win/loss record (there is no alliance win/loss record as such now).
This isn’t a defence of how war matchmaking occurs, but an attempt to explain what I understand about it.
Additional: I wrote the above the other day but it took a while to formulate. But since the discussion has continued…
Hopefully my attempt above will go some way to explaining the war score aspect and perceived discrepancies.
Though you seem to focus on the defence teams’ TP on the field of battle a bit too much. There are level 30-something players with TP4200+ raid defence teams comprised of mostly 5* heroes. What about that level 70 player with only a TP2000 raid defence team? Do not let TP mislead you. Look to other things like the tank’s DEF stat, and the points worth of the team. Also consider their specials and whether your heroes can counteract them. There’s a lot more than just TP to consider in attack.
If you truly believe it’s a bug, then submit information to the bug thread or file a direct ticket to support. While I do believe the war matchmaking algorithm isn’t necessarily satisfactory, it doesn’t mean it is broken.