Dear All,
I have updated and completely revised my old manasheet. Here you can get the latest version. Both ninja and magic heroes have been added. With the ninjas there is often an inaccuracy in the calculation, which I have fixed here.
The blue table is to be understood as an overview of all currently possible breakpoints. Just look at which mana bonuses you can get together with a hero (yellow table) and then compare with the blue table whether and where you can reach breakpoints.
It is always highly recommended to prioritise the breakpoints on 9 and 6 tiles respectively. It makes a huge difference if you get heroes with average mana speed loaded with 9 tiles instead of 10. This often decides the outcome of a fight.
With the green table, I have created an extra table for the Rush Mode. This is about the fact that you cannot use mana troops in the weekly 3* and 4* tournaments, for example. Heroes who can be accelerated to 6 tiles in rush mode even without such a troop become particularly valuable and practical. Also with the 5* heroes, it can be interesting when you can equip a hero with, for example, a ninja or crit troop and still get him accelerated down to 6 tiles. A big thank you goes here to Drea, who was the first to come up with the idea of creating such an overview for Rush Mode.
The manasheet is explicitly designed for calculation in attack only. Contrary to what is communicated in many places, the calculation in defence does NOT work analogously according to the motto: half a mana unit per incoming stone and a whole mana unit per turn. This is demonstrably wrong! Rather, heroes in defence apparently need a little more mana to fully charge than in attack. Furthermore, the amount of mana generated in the defence team is reduced per incoming follow-up combo. My theory is that it should be about 20% more mana for defence teams to generate, based on the above scheme of “half a mana unit per incoming stone and a whole mana unit per turn”, and that the mana generated is also reduced by about 20% per follow-up combo. This would mathematically result in only a fraction of the previously generated mana being generated after the 10th follow-up combo. Please note, however, that this is pure speculation. It is currently beyond my time to test extensively and try to get to the bottom of this.
Nevertheless, the sheet now available on the attack is quite complete and comprehensive. I hope it helps many players in their in-game planning and contributes to a better understanding of the mechanics of the game. If you have any questions, post them in the comments. I look forward to it and will try answer them all to the best of my ability.
Many greetings
Faeldin
Edit 1 (2022-03-23 and again 2022-03-27): I have corrected the figures for the first stage of magic charging. Thanks to @Liam_K and later @Zloykot for the hint! 
Edit 2: (2022-03-24): I have added a German version (Deutsche Version) of my Mana sheet below.
Edit 3 (2022-04-04): I have now added the already verified numbers for the magic heroes. All numbers that have not yet been verified have been removed for the time being to be sure not to spread false data. It remains to be checked whether the calculation of the magic heroes must be based on a different formula than the known formula or whether there is another solution to the problem. At the moment, unfortunately, the experimentally checked values at charge 2 do not come to a common denominator. Another edit: Furthermore, a value at Charge 3 of the ninja heroes was checked and corrected by me. These actually charge with a mana bonus of only 5% and 14 tiles on Charge 3.
Deutsche Version des Manasheets:
we don’t really know the exact decay rate nor the base mana units per tile…
