Mega Delta Worksheet

[Deck Knight]

Way back when we had the Mega Evolution thread before Crucibelle, we discussed a concept called Mega Deltas, or how much a mega evolution improves the base Pokemon competitively. I wanted to find a way to quantify this as it could help us in future projects. Note I say improve rather than change because, as you will notice, Mega Garchomp barely improves the base because the speed drop is detrimental, so if we were to have a CAP with that as a feature of the Mega, it would need to compensate elsewhere.

Methodology and Purpose:
I used the Smogon damage calculator to test the various Mega Pokemon against a Standard Opponent, and also tested them defensively against a Standard Attacker and Standard Special Attacker.

In this way the ratings I derived would be based on percentages in realistic, relevant battle settings.

Link: https://onedrive.live.com/view.aspx?resid=1E6751FA2224989C!415&app=Excel&authkey=!AOPGXLxaqh7Mwe4

Caveats / Explanation:
Offensive ratings (SPAF/SSAF):
Offensive ratings are for maximum investment without nature and based on the Pokemon's strongest legal attack (factored for accuracy), including abilities, but not including natures or items. So if you're wondering why Mega Charizard Y rivals Mega Charizard X physically and Mega Gardevoir rivals Mega Lopunny physically, it's because Drought Flare Blitz and Pixilate Double-Edge are disgustingly powerful, perfectly legal options available to those Pokemon. While I don't expect Mono-Attacking Return Mega Gardevoir to be beating the life out of Chansey any time soon, it would catch it off guard and regular Gardevoir can't do anything anything like that physically.

Offensive ratings are relatively accurate depictions of the percentage damage the Pokemon would do on average to a minimum investment (8 HP EVs to simulate 4 Def/SpD EVs) 100 HP/Def/SpD Pokemon, a Pokemon of relatively middling bulk, set roughly at Manaphy or offensive Jirachi levels. If a Pokemon's strongest move is inaccurate the BP was adjusted by accuracy (so Mega Lopunny HJK is set at 117 Base Power). The offensive ratings tell you (roughly) how well offensive Megas attack, not how well defensive Megas attack.

Defensive Ratings (PEND/SEND/WRM):
Defensive ratings are for minimum investment (8 HP EVs to simulate 4 Def/SpD EVs) so if you're wondering why Mega Sableye has a lower Physical Endurance (PEND) rating than Mega Pinsir, it's because maxing Sableye's HP greatly increases it's overall defenses whereas, in a vacuum, 65/120 is better than 50/125. Defensive ratings do consider abilities like Intimidate on the Base Form or Mega, but only for that Pokemon (so Gyarados's PEND rating is better than Mega Gyarados's).

There was no efficient way to be able to differentiate Defensive Megas from Offensive Megas in a way that wouldn't turn calculations into a nightmare. The object of the sheet is to identify change from Base to Mega. Mega Sableye's PEND difference is huge, Mega Pinsir's is small. TL;DR Defensive Ratings tell you how well offensive Megas take attacks, not how well defensive ones take attacks.

The Weakness Resistance Modifier (WRM) considers the Pokemon's overall type resistance. Types that remain the same have no difference, types that change have that change reflected in their defensive ratings. WRM uses the sum of all the Pokemon's weaknesses and resistances and divides by 18. It factors in Abilities like Filter which change weaknesses from 2x to 1.5x.

Standard Attackers:
The Standard Attackers chosen for the defensive ratings were Jolly Rocky Helmet Garchomp Earthquake (STAB Neutral Max Atk) and Naive/Timid Assault Vest Tornadus Hurricane (STAB Neutral Max SpA).
These seemed to be realistic, relevant metagame threats whose damage percentages could form an effective baseline for univested strength. If these particular attacks easily 2HKO an univested Base but barely 2HKO an uninvested Mega, that defensive difference is competitively relevant.

Speed Significance:
This was a multiplier I derived for offensive and defensive deltas based mostly on important speed tiers. I used Base 30, 70, 80, 100, 105, 110, 120, 135, and 145. If the Pokemon shifts up or down one of these speed tiers their speed significance factor is likewise adjusted, based at 1 and then +/- 0.5 on the multiplier for each shift. Shifts are inclusive of tying that speed so Mega Pinsir's multiplier is 2 (1 Base + shifting above 100 and at 105 speed tier, 0.5 x 2 = 1, 1+1=2). Mega Alakazam and Mega Aerodactyl's are also 2 because they shift up above 135 and 145. Since Mega Camerupt, Mega Sableye, Mega Garchomp, and Mega Ampharos all lose speed their Speed Significance multiplier is 0.5.

Deltas:

• O-Delta (Offensive Delta) averages the Pokemon's Differences in SPAF, SSAF, and multiplies by Speed Significance.
• D-Delta (Defensive Delta) averages the Pokemon's Differences in PEND, SEND, multiplies by WRM difference, and also multiplies by half the Speed Significance. This is because faster or slower Pokemon may have to take / will avoid additional hits. It doesn't effect Pokemon with no defensive differences between Base and Mega because those Pokemon don't have a net difference in defenses [Thanks, Math!].
  • I don't know whether it's relevant to adjust for this or not. In theory if you can take a hit when you Mega Evolve and be faster the next turn, your defense has improved (take less hits). In reality though, the three Megas this applies to (Absol, Beedrill, Glalie) have the defenses of a wet sack, and in fact would be KO'd by a Life Orb attack from either of the Standard Attackers.
• Total-Delta is the sum of O-Delta and D-Delta.

Conclusions:

Crucibelle fell well within normal parameters for a Mega Delta, it's change is actually less aggressive than the average (41.9 compared to 48.0), but not by much. Compared to the average it has a smaller offensive delta (23.4 compared to 35.1) and a larger defensive delta (18.5 compared to 13.2).

The Standard Deviation on Megas is pretty high (31.8). If you were to set a CAP boundary on Mega Deltas at 1 Standard Deviation, you could go as low as Mega Ampharos (19.5) or as high as Mega Gardevoir (78.3). Mega Garchomp (14.4) would be too little competitive improvement, even though 102 and 92 speeds play entirely differently.

Credits:

I worked on most of the grinding, but I want to thank Imanalt for pointing me in a better direction for factoring in Speed shifts. Whether this was an effective solution I'll leave up for critique, but I think it captures competitive efficacy well enough, and gives us a realistic window into how much improvement is "average" for Gamefreak created Mega Pokemon.

Fun Facts:
The 3 Lowest Deltas are Mega Gyarados (10.6), Mega Slowbro (12.5), and Mega Garchomp(14.4). Since this is an improvement score this means the originals are pretty good themselves.
The 3 Highest Deltas are Mega Pinsir (128.6), Mega Lucario (124.6), and Mega Sharpedo (119.4). Mega Sharpedo gets about 60% of its boost from going from membrane defenses all the way to paper, the rest is Strong Jaws Crunch.

 

[LightningLord2]

Is it meaningful to compare this Mega Delta with the Delta that goes with equipping a different item that augments stats? That is, how much a Pokémon improves when wearing a Choice Item, Life Orb, Eviolite (no mon can choose between Eviolite and a mega stone so far) or an Assault Vest. This is probably relevant because there's only marginal advantages for Blaziken to use Blazikenite over Life Orb (mostly 10 more base speed), which could apply to future Mega CAPs as well.

 

[Deck Knight]

Is it meaningful to compare this Mega Delta with the Delta that goes with equipping a different item that augments stats? That is, how much a Pokémon improves when wearing a Choice Item, Life Orb, Eviolite (no mon can choose between Eviolite and a mega stone so far) or an Assault Vest. This is probably relevant because there's only marginal advantages for Blaziken to use Blazikenite over Life Orb (mostly 10 more base speed), which could apply to future Mega CAPs as well.

Not the Mega Delta itself, but the derivatives SPAF, SSAF, PEND, and SEND are percentages, so a Life Orb would Increase the the ratings by 30% (x1.3). Choice Band/Specs would increase them by 50%. Assault Vest would Decrease SEND by 50% (x0.667). Positive Natures would increase SPAF/SSAF totals by another 10% (x1.1), and Decrease PEND/SEND by 10% (x0.9). Negative Defensive Natures Increase PEND/SEND by 10% (x1.1)

EX: Abomasnow's SPAF is 49.05, if you do the same calculation with Life Orb it's 63.8. Band is 73.7.
Life Orb is higher than Mega Abomasnow's SPAF rating (63.25) by a small margin, but obviously Mega Abomasnow doesn't lose health when it attacks.

Assault Vest decreases SEND (remember, lower ratings mean LESS damage), it would lower Abomasnow's SEND from 67.75 to 45.45, which is lower than Mega Aboasnow's but obviously then it can't use any non-damaging moves.

The next thing I can do is work on a calculator for the sheet using these standard ratings, but I wanted to get some initial feedback on whether people think these ratings are useful or not. In many cases a Life Orb attacker has more damage output than the Mega Evolution, and an Assault Vest prevents more special defensive damage than a Mega Evolution, but since both those items have drawbacks and only one can be attached at a time, it made more sense to just say "Mega Abomasnow is this much more improved from Base Abomasnow" instead of "Mega Abomasnow is X improvement better than LO Abomasnow (value) and X improvement better than AV Abomasnow (different value).

 

[Earth Traveler]

I was looking at your chart, and, while it does have some helpful information regarding how Pokemon get changed by mega-evolving (we can now quantify how good/bad some Megas are), I was wondering about one thing (or maybe six things, to be more exact). If this chart is supposed to help us make future CAPmons, which are currently designed for the OU meta, why did you choose to include the Ubers Megas, such as Mega-Blaziken, Mega-Mewtwo's, etc.? These megas could be throwing off averages in my opinion, so... I'm not sure why they're there.
Sorry if this is a dumb question.

 

[Deck Knight]

I was looking at your chart, and, while it does have some helpful information regarding how Pokemon get changed by mega-evolving (we can now quantify how good/bad some Megas are), I was wondering about one thing (or maybe six things, to be more exact). If this chart is supposed to help us make future CAPmons, which are currently designed for the OU meta, why did you choose to include the Ubers Megas, such as Mega-Blaziken, Mega-Mewtwo's, etc.? These megas could be throwing off averages in my opinion, so... I'm not sure why they're there.
Sorry if this is a dumb question.

Ironically, the Mewtwos have a moderating effect on averages because the Megas score improvement. Mewtwo already does tons of damage and takes hits pretty well so relatively speaking the damage differentials on its Megas are not that high. The same with Mega Blaziken really, it's got +40 Atk, yes, but so did Altaria, and Altaria gets boosted by Pixilate.

The only one I didn't include is Mega Rayquaza or the Primals. I suppose I could include Mega-Ray even though it would have a pretty high score (also I'd feel obligated to find some way to factor in that thing can still use items lol), but since the Primals don't even have a Mega Evolution turn I don't think they're valid comparisons. They're essentially entirely different Pokemon.

 

[QxC4eva]

I wasn't going to post, but then I saw this

The next thing I can do is work on a calculator for the sheet using these standard ratings

so I have to post >_>

I don't think they're good ratings, and I hope they don't become standard. The best way I know to calculate mega improvement is the difference between raw sweepiness and tankiness (RPS, RSS, RPT and RST) of the mega and base. X-Act has done all the work on those ratings, they're objective, and there's few reason why you'd want to make up your own ones for a similar purpose like this.

Good ratings are objective measurements, based on derived formula. Yours is the opposite of that. For example, Antar's Viability Ceiling measures the highest GXE of a player who uses the pokemon, X-Act's Raw Physical Sweepiness measures the average number of half-turns a physical attacker needs to KO each pokemon, and Dragontamer's Blisscents measures the damage percentage a special attack will inflict on a Blissey. To me these are all good ratings, and I have huge respect for the great minds behind them.

With that said, here's my take on each of yours.

• SPAF/SSAF: There's very little difference whether you pick a target with 100/100 bulk or something else for this benchmark. As long as the target is at level 100, the results will be pretty much proportional. Why not just stick to how X-Act does it?
  
  
  
  will get you like the same results. :\
  Hide

  Let's try it on some pokemon...
  
  Lopunny: (76*2+99)*102*1.5/1030 = 37.3
  Mega Gallade: (165*2+99)*120*1.5/1030 = 75.0
  Mega Manectric: (135*2+99)*90*1.5/1030 = 48.4
  
  Now look at your chart. Do the numbers look familiar?
• PEND/SEND: Similar to above, you're better off following how X-Act does it in RPT and RST. Oh, and calling it "endurance" is misleading. Endurance means long lasting, which essentially are mons that can recover HP which has nothing to do with this.
• WRM: Effectiveness roughly follows a logarithmic scale. So if you take the arithmetic average, you won't get the average, you just get some rubbish!!
• Speed Significance: Why start at 1? Why increment by 0.5? Where did these magic numbers come from, and what are you objectively trying to measure?
• O-Delta: Why average the SPAF and SSAF difference? Why multiply by speed significance? Where did this magic formula come from, and what are you objectively trying to measure?
• D-Delta: Why multiply by WRM difference? Why multiply by 0.5 speed significance? Why not multiply by another number like 0.4 instead? Where did this magic formula come from, and what are you objectively trying to measure?

Also, please cut down on the acronyms! </3 It's very confusing to read. I don't think your ratings are good enough to deserve their own acronyms yet. For now, a good rule of thumb is less code, more english.

 

[Deck Knight]

For reference I'm not married to any particular formula, however the reason I only pulled speed factors from the BSR spreadsheet is that, broadly speaking, BSR's greatest weakness is it doesn't account for two very relevant attributes: Move Base Power and Abilities. Instead I opted for damage averages from the damage calculator.

If you just used BSR, the only difference between Charizard and MegaZard X would be whatever difference the attack increase adds to its sweepiness. But there's a problem- MegaZard X is a lot "sweepier" than its BSR rating would lead you to believe. Statistically it's less than Garchomp because 130/102 > 130/100, but ZardX hits a lot harder than Garchomp in practical battling scenarios, unless Chomp is using Life Orb (which has a negative impact on tankiness.) This is even more potent when you consider something like Mega Pinsir. The BSR Sweepiness differences in no way accurately depicts the fact Pinsir went from having its strongest move change from 80 Base Power STAB X-Scissor to 159.6 Base Power STAB Double-Edge.

I'm not nearly as mathematically skilled as X-Act, so I did the best I could based on the skills I do have, which is to take objective anchoring points and use them as a reference. The speed significance is too subjective in my opinion and I'm happy to replace it, but failing a model which tells me specifically of it's more important to go from 85 to 105 than 105 to 135, I opted for a cruder method.

Anyway, suggestions are welcome, I'd just like to empasize that I'm trying to gauge improvement and not overall Mega strength, i.e. I'm looking for how much Mega Pinsir changes Pinsir not whether Mega Pinsir is better than MegaZard X.

 

[QxC4eva]

BSR's greatest weakness is it doesn't account for two very relevant attributes: Move Base Power and Abilities.

Moves and abilities can be injected into BSR's formula, you know. Did you see how I added base power to it on my last post?
All you really need to do is understand the math, then you can fully customize it. Don't just assume you can't :P

Instead I opted for damage averages from the damage calculator.

You're treating the damage calculator like a black box. Fun Fact: we all know what's inside it! Just grab that formula and chuck it into your math. Notice how certain variables will cancel each other out (such as the Manaphy you chose to do the benchmark on) You'll end up with simpler calculations that are just as relevant as the numbers you get from the damage calculator. In my last post, I showed you how the numbers on your chart can fit into one simple formula.

The speed significance is too subjective in my opinion and I'm happy to replace it

• Your Speed Significance is a subjective version of Speed Factor
• Your Endurance rating is outclassed by Tankiness
• The way you combine speed and attack is inferior to Sweepiness
• For what you're trying to do in WRM, I think you should read up Defensive Type Rating
  
  Hide

  Unsurprisingly, we could also incorporate this Type Defense with the Physical and Special Tankiness in the Base Stats Ratings

  ^ so yup, looks like you can inject this into BSR as well
• To find the Deltas, take a look at the math behind Offense Defense Balance (ODB). For Offensive Delta, divide the mega's sweepiness by the base's. Then do the same thing for tankiness to get Defensive Delta.
• Total Delta is the same idea as X-Act's Overall Rating formula - you multiply the deltas together.

I think "improvement" is a better word than delta though.

 

[Deck Knight]

Updated the sheet. Being able to plug in Base Stats is very helpful.

That said, I'm wondering where the divisor 1030 came from. Can't find it in either the damage article or X-Act's original.

The other problem with Type Defense Ratings is that the weighting X-Act put on them is now obsolete. It was made before BW was even released, so a lot of the behind the scenes weighting is now irrelevant. XY added Fairy Types which I can safely assume dumped Fighting's damage output weighting into the gutter all on its own. Since I don't know specifically how X-Act weighted them I'm open to suggestions on how to do something like Type Defense competently. Also one of his sentences was "I'm also weighing the fact that Earthquake is used more than, say, Brave Bird." Which may or may not still be true but the existence of Talonflame in OU makes it less true than it was in July of 2010.

After going through the model you suggested QxC4eva there's a lot I like about it. The first is that if we just strictly use the BSR ratings, the improvement calculations are easy and can just be an add-on to that sheet. My concerns with doing that are the following:

1. It still doesn't mathematically account for abilities impacting Sweepiness. You can't just multiply ZardX's Physivsl Sweepiness Rating by 1.33x and get the proper rating for it. Here's the formula for Physical Sweepiness with MegaZard X's Base Stats and Speed Factor:

=(((((130*2)+36)*((((130*2)+36)*(698/776))+415))/((((130*2)+36)*(1-(698/776)))+415)))/2.3116+6.43858

In X-Act's original you found the normalized Base Stats by taking the Base and adding 18. Looks like the formula has changed though. I assume the last divisor is for weighting purposes but nonetheless, I can't figure out where, mathematically I'm supposed to add in the fact ZardX is going to have 1.33x power on its common contact attacks as a physically sweeper.

The same goes for trying to factor in Adaptability, Sheer Force, -Ate Abilities and Drought. These abilities significantly and consistently improve damage output and not accounting for them does not accurately portray the level of improvement we're trying to measure.

I'm more than happy to do the work necessary for the calculators, I just need to figure out what to factor in and where to factor it.

2. Same issue with the Weakness/Resistance Modifier and/or Type Defense. Mathematically I'm in a dark zone as to how it was weighted, so as crude as "Sum W/R mulitpliers and divide by 18 for number of types" is, short of a better alternative its a representation a layman can understand and it affects so few Megas significantly it's a decent way to capture it.

 

[Ignus]

I really like the idea behind this. Good topic for analysis!

That said, I'm going to echo QxC on the formula's accuracy - stick to X-Act's for dealing with how strong a Pokemon is offensively. It works really well.

As for the type/defensive rating, I've been working on (and losing the code due to hard drive failures) an objective formula for that as well - It won't be finished soon because I'm slow, but when I do we can see if it's worth implementing here. Until then, you could try this:

Find the damage of each Pokemon's highest/most common base power STABs in OU, multiply that by usage. Do that for every pokemon of the same type. Divide by the total usage. Do that for every type. Next, multiply each type's value by the Typing Modifier of the defending Pokemon. After that you can just add all of those values up for one big number that represents how strong the typing is.

Here's the 'mathy' formula for that:

The great thing about this formula is that it's really easy to turn into usage weighted 'average damage taken' - this equation is probably the closest thing we can do to an actual mean damage taken without looking at each individual battle and averaging from there (which is the thing I'm doing. My shit keeps breaking. One day I'll finish).
For the full 'defensive' formula, you just take the previous equation and rather than just adding them together, you take an average of each type's value and multiply that by the targets defense and divide it by target HP. This will give you a percentage that literally represent how much damage the Pokemon will take on average from opponents. Kinda cool.

That's all I have until I figure out how to make my crap stop breaking

 

[QxC4eva]

Now we're talking :D

I'm wondering where the divisor 1030 came from. Can't find it in either the damage article or X-Act's original.

1030 came from your Manaphy benchmark. If you picked something else like Phione (80/80 bulk) the divisor will be lower to something like 750. Basically, whichever mon you pick for the benchmark will act as a divisor in the formula. I'll spare you the math as to why :P

The other problem with Type Defense Ratings is that the weighting X-Act put on them is now obsolete. It was made before BW was even released, so a lot of the behind the scenes weighting is now irrelevant. XY added Fairy Types which I can safely assume dumped Fighting's damage output weighting into the gutter all on its own.

Okay. The reason I suggested reading X-Act's stuff is so you get an idea how he does it. I agree it's kinda obsolete but the concepts are still the same. So yeah. You can't just copy his work, you have to understand how he derived it and tweak it so it's relevant for what you're trying to do. There's no free ride here.

1. It still doesn't mathematically account for abilities impacting Sweepiness.

>_> Since you're not really getting it, I guess I'll walk you through.....

1. How to find SPAF

First off here's the damage formula:

We get rid of the + 2 and roundings cause they tend to overcomplicate things, and barely have an effect on the results anyway.

We know some values already. Level = 100, Atk EVs = 252, Def EVs = 0, IVs = 31, Natures = 1 (neutral), and Damage Roll = 0.925. Sub them in:

and simplify:

SPAF measures the percentage of HP lost, but since your chart ignores the % sign we multiply by 100.

Sub in the Damage and HP formulas:

You chosen a 8 HP Manaphy as the target, so the Base HP = 100, Base Def = 100, HP IVs = 31, HP EVs = 8, and Level = 100.

Simplify and.....does it look familiar??

I got 1042 this time instead of 1030, but anything from say 950 to 1050 should be OK!


Let's test the formula with Zard Y in the sun: (2*159+99)/1042*110*1.5*1.5*0.5 = 49.5
252 SpA Mega Charizard Y Fire Blast vs. 8 HP / 0 SpD Manaphy in Sun: 156-185 (45.4 - 53.9%)

And the Zard X you've been losing sleep over: (2*130+99)/1042*120*1.5*1.33 = 82.5
252 Atk Tough Claws Mega Charizard X Outrage vs. 8 HP / 0 Def Manaphy: 255-301 (74.3 - 87.7%)

Looks good I think! For simplicity, I suggest cross out the divisor (scrap the Manaphy benchmark) it's not like you need it to find the mega delta anyway... :\

2. How to find Sweepiness

We measure Sweepiness as the average damage a pokemon can do before it gets hit.
So if you think about it, there's two ways to increase a pokemon's sweepiness:

• Increase attack, so the damage output is higher
• Increase speed, so it gets to claim damage on more enemies that it outspeeds

To do that, X-Act came up with a neat idea called "half-turn ratio" (RHT). Half-turn is when ONE pokemon makes a move, instead of 'full turn' when BOTH sides make a move. Got it? It's just a fancy word, nothing scary :S Anyway, the ratio is the average half-turns a pokemon makes per half-turn the opponent makes:

The faster you are, the more likely you go first in a random matchup. So we need to find the probability of being faster, and Speed Factor (SF) does exactly that:

I put a ? because in pokemon, that's wrong - both players get to move on each turn. So you can't be more than 1 turn ahead or behind no matter how big the speed difference is. E.g. a pokemon with 0.8 SF doesn't move twice as often against something with 0.4 SF. To model it, we add the average number of turns per KO (t) on both probabilities:

I'm guessing t is about 1.5 for OU but maybe you can decide! Just for reference, it should only average the offensive part of the meta. Sweepiness don't apply for mons that don't sweep, so count out the stall, balance mons etc. The ones that do matter (sweepers or anything HO) rarely survive more than 2 hits from each other so I set mine to 1.5.

Sweepiness is the half-turn ratio weighted by damage output (which we all know is SPAF)

Sub in SPAF formula and we're done! I got rid of the 1042 divisor because who needs that.

where:
ATKbase = base attack stat of the attacker
BP = base power of the attack
SF = speed factor of the attacker (0 to 1)
Modifiers = multipliers like effectiveness, crit, STAB, abilities, items, weather, etc.
t = average turns per KO in the offensive metagame (1.5?)


There you go!! I tested the swepiness formula on a few mons who can mega:

• Audino: 28221 --> 28221 (+0%)
• Beedrill: 37754 --> 105553 (+180%)
• Garchomp: 95182 --> 104881 (+10%)
• Lopunny: 66248 --> 92727 (+40%)
• Sableye: 30672 --> 24531 (-20%)

3. How to find Tankiness

Remember the SPAF formula before we put the numbers in for Manaphy?

well Tankiness is the opposite that, so we flip over the fraction:

You picked Garchomp's EQ for the benchmark, so Base Atk = 130, BP = 100 and Modifiers = 1.5 (STAB). The defender is level 100 with 31 IVs and 8 EVs for HP.
Shove it all in:

Simplify.

Done, but that divisor looks so stupid!! You should get rid of it imo >_>


It should match any benchmark quite well. Pick a random mon like Phione: ((2*80+143)*(80+18)/2092000)^-1 = 70.5
252 Atk Garchomp Earthquake vs. 8 HP / 0 Def Phione: 196-232 (64.6 - 76.5%)

4. How to find the Deltas

Determine the scale factor, then subtract 100% to get the % net change

Testing the Overall Delta:

• Mega Audino: (28221*26177*50256*50256/(28221*24901*36296*36296))^0.25-1 = +19.1%
• Mega Beedrill: (105553*38392*15834*26754/(37754*28772*15834*26754))^0.25-1 = +39.0%
• Mega Garchomp: (104881*78965*47747*40567/(95182*66952*40567*36977))^0.25-1 = +13.8%
• Mega Lopunny: (93230*45898*30576*31122/(66248*42027*27846*31122))^0.25-1 = +14.0%
• Mega Sableye: (24531*20128*34749*32319/(30672*23145*22599*20169))^0.25-1 = +14.4%

2. Same issue with the Weakness/Resistance Modifier and/or Type Defense. Mathematically I'm in a dark zone as to how it was weighted, so as crude as "Sum W/R mulitpliers and divide by 18 for number of types" is, short of a better alternative its a representation a layman can understand and it affects so few Megas significantly it's a decent way to capture it.

Hahah fair enough :P but I gave you flak for a few reasons:

• Your math is wrong. The effectiveness multiplier sequence goes: 0, 0.25, 0.5, 1, 2, 4 which fits more as a geometric sequence than arithmetic one. Here's how to find a geometric mean. Immunity will mess things up but from experience, a decent hack I do is set it to 0.0625
• You went ahead and calculated the WRM difference. WRM is a multiplier! You're meant to multiply it onto SP/SAF to offset their values, not subtract it from another WRM then do something weird with it.... :|
• It's not consistent with your other ratings. Speed Factor sorts every pokemon in speed order first then do stuff with it. So for here, you should sort every attacking move into their type first then do stuff with it. Even better weight them by usage (like the formula by Ignus shown above) if you don't mind doing the same thing to Speed Factor as well. Anyway it's no biggie, looks you're just getting the average multiplier which is fine, but refer to point 1 on how to do it correctly -.o