GamePress

Rounding issues with quick attacks

I read this remark in a posting by Data10, and it really opened my eyes:

"For example, I powered up my attacking B/SE Tyranitar to level 33. I chose level 33 because that will allow me to fight up to level ~35 Blisseys with Bite rounding to 5 damage instead of 4, which is the strongest Blisseys I've faced in my local gym meta. And getting that 5 damage from Bite instead of 4 will reduce battle time by 10-15 seconds, which is very important against Blissey."

Normally, i ignore those rounding issues, because they are hard to calculate with, but they seem to be quite relevant.

The example given above is good - rounding in Battle gives whole numbers of damage delt for each attack. If calculation gives 4.4 damage, you would deal 4 damage, if calculation gives 4.6 damage, you would deal 5 damage, each time a quick attack deals damage. In this example, difference is huge - 25% - as much as being super effective or not.

[Edit]
Correction - as in damage formula "floor" is used instead of "round", the example should have been 4.9 versus 5.1. Thanks to blue-333
[/Edit]

This effect is the more important, the smaller and faster an attack is. It can be completely ignored for charged attacks. It should be concidered for quick attacks with up to 9 damage, above the effect is smaller.

How can we use this in praxis? As damage is calculated out of level dependend attackers ATT-Stat, defenders DEF-Stat as well as SE and STAB, it depends highly on the matchup. If you have a generic attacker as lick-snorlax, you have so many different matchups, that this effect is middeled out over many battles. The interesting part comes with repeating specific battles - as Data10 writes, Tyrannitar versus level 35 Blissey. Another interesting matchup is venussaur - vaporeon. In the range between level 30 and 34 Venussaur and Vaporeon, either 6 or 7 damage are delt with each Vip Wine. So - for me - it makes absolutely sense to raise my Venussaur to level 34, as it is likely to be noticeable better in its favorite matchups.

An easy way to test those specific matchups is PokeBattler. After simulation, you can look into the battle log and see how much damage is delt with your quick attacks. Then you can vary attackers and defenders size and maybe you notice such a step, making you rethink leveling beyond 30 or not.

Regards BB

Asked by Bruno Brezel8 years 3 months ago
Report

Answers

4.4 and 4.6 all does 5 damage.
The damage formula is: Floor[0.5Power(ATK/DEF)STABEffective]+1
However, not: Round[0.5Power(ATK/DEF)STABEffective]+1

Up
0
Down

so take 4.9 and 5.1. Thanks for correction.

This mistake does not influence the content of my statement, it just changes the position of those interesting breakpoints.

Up
0
Down

I've seen this mentioned before on reddit/silph on why some ultra high level trainers level some or all their key attackers to level 39 (40 if possible).

Getting that extra 1 damage, as you note can make battles insanely faster and also why in some videos, some moves maybe better than others (my guess is this was where the FunkySho video of a Bite/Fireblast Arcanine was faster than a Fire Fang one in a very early video)...

Overall, I think when you get to the upper tier levels and you have more stardust than mons to really power up (already in 10 gyms, etc...), you probably SHOULD power up a few attackers over 30 to save time.

The best candidate in my mind is Machamp since the Blissey and Snorlax fights are still time consuming while vs. things like Dragonite, a level 30 Jynx (FB/A) can do in less than 20 seconds (pokebattler sim)...

Up
0
Down

I have run some simulations, see
https://pokemongo.gamepress.gg/q-a/taking-advantage-rounding-issues

The best candidate in my mind is Machamp since the Blissey and Snorlax fights are still time consuming while vs. things like Dragonite, a level 30 Jynx (FB/A) can do in less than 20 seconds (pokebattler sim)...

the Machamp-Blissey-Matchup seems not to profit from rounding issues, at least not in my simulation.

Up
0
Down

Here's a Machamp vs. Blissey from Pokebattler:

Lvl 30 Machamp (C/DP) vs. Lvl 39 Blissey (ZH/DG) maxed IVs for both, beats Blissey in 64.8 secs and Counter does 9 damage:
https://www.pokebattler.com/fights/attackers/MACHAMP/quickMoves/COUNTER_FAST/cinMoves/DYNAMIC_PUNCH/levels/30/ivs/FFF/defenders/BLISSEY/quickMoves/ZEN_HEADBUTT_FAST/cinMoves/DAZZLING_GLEAM/levels/39/ivs/FFF/strategies/DODGE_ALL/DEFENSE

You take Machamp to Lvl 32 (C/DP) vs. Lvl 39 Blissey (ZH/DG) maxed IVs for both, beats Blissey in 61.1 secs and Counter now does 10 damage:
https://www.pokebattler.com/fights/attackers/MACHAMP/quickMoves/COUNTER_FAST/cinMoves/DYNAMIC_PUNCH/levels/32/ivs/FFF/defenders/BLISSEY/quickMoves/ZEN_HEADBUTT_FAST/cinMoves/DAZZLING_GLEAM/levels/39/ivs/FFF/strategies/DODGE_ALL/DEFENSE

At level 39, the Machamp is STILL at 61.1secs and Counter still does 10 damage so no more real benefit:
https://www.pokebattler.com/fights/attackers/MACHAMP/quickMoves/COUNTER_FAST/cinMoves/DYNAMIC_PUNCH/levels/39/ivs/FFF/defenders/BLISSEY/quickMoves/ZEN_HEADBUTT_FAST/cinMoves/DAZZLING_GLEAM/levels/39/ivs/FFF/strategies/DODGE_ALL/DEFENSE

This is probably because Dynamic Punch overkills Blissey as the last move...

Up
0
Down

So my C/CC Machmap will soon upgrade from lvl 27 to lvl 32 :). (btw, at lvl 27 C does the same 9 damange).

Up
0
Down