I’m going to start off with a pair of disclaimers:
1) In any scientific inquiry if theory and experimental results conflict then it is the results that are correct. If your method is working for you, and working noticeably better than the more conventional methods then by all means continue. As part of this I should note that my only evidence about your method comes from your description so if I’m missing some nuance or detail that’s why.
2) I’m going to be going to be talking about maindecking sideboard cards and preboarding in my next article, so if you think I’m missing some detail about those it’s because I’m saving it for Tuesday.

In theory there is validity to your method. Playing broadly effective maindeck sideboard cards can be very effective. Using data analysis and spreadsheeting as you describe is fairly common (I think Frank Karsten has written about something similar a few times) and it can be effective. The problem is that it is very easy to get too “in the weeds” and lose sight of the big picture. It sounds like you’re playing Affinity so let me ask you how much your method is impacting your goldfish? Maindecking sideboard cards in the numbers you’re describing in aggressive decks is dangerous because you can dilute your gameplan too much. If you’re bringing in and maindecking a lot of cards I have to assume that you’re also taking out the support cards for the maindeck plan. Too much of that and suddenly your maindeck plan doesn’t work anymore. If you’re removing Affinity’s usual flex cards (almost always Galvanic Blast these days) I have to ask whether on average the ability to goldfish faster would have been better than a card that buys you time in some matchups.

I cannot see your data to give you a definitive numbers based explanation, but I can say that the questions you have to answer are about overspecialization and opportunity costs. Does playing more specialized cards that are targeted at certain matchups but have some play elsewhere give better percentage across the board than playing a card that has the same value regardless of matchup? Say you have one card whose value, on a scale of 1-10, is 8 against certain decks and 3 against everything else. The card you are thinking of replacing for is consistently a 5.5 regardless of matchup. The question you have to answer is whether you will get the 8 enough, with its +2.5 payoff, enough to justify the -2.5. I don’t know the answer for you, but you will need to answer it for yourself if you want to keep using your method.

As for your specific method I have questions about your procedure that makes me question the validity of your results. As described your methodology is too complex and inconsistent. Why are you adding value on 10% and subtracting value on 8%, and why is that value equal to the percentage? That appears to be too complex and arbitrary for really valid analysis. I would suggest, based on my education and experience, that you should choose a baseline percentage, for my purposes here the middle of Tier 1’s range (5%), set that equal to zero and base your scoring off of that number. In my system, using the data from our Top Decks page as I’m typing this, Jund would be 3.7 (8.7%-5%)while Affinity would be -.4 (4.6%-5%). It doesn’t matter where you put the baseline (the dividing line between Tier’s 1 and 2 may be better) as long as you have one and use it consistently. This may give you a very different picture of the meta and your sideboarding decisions than your current method but it would also be more consistent numerically.

I hope this helps, and if you have more questions either ask here or see if I cover them on Tuesday.

What you’re saying is true, but that’s also generally true of any completely fair midrange deck regardless of format. It’s arguably GBx’s greatest strength and the reason to play the deck. You’re definitely not on the deck because its 50% vs almost everything. In my experience the best way to cripple Jund in particular has been mana denial. Spreading Seas and Tectonic Edge really put the screws to the deck though they’re not the blowout wins that you may want. Frankly, since it sounds like you’re on a control deck I think that you’re best bet is Sun Titan recurring value creatures and Tec Edge to simply grind Jund out rather than looking for bullets, since they really don’t exist.

I feel like the idea behind the rule is to make sure you have some form of graveyard hate, and dredge just makes a good poster boy for the concept. As you know, modern does have a lot of graveyard interactions between goyf, snapcaster, delve, and many others. It only makes sense to pack hate for these interactions when such good hate exists.

Well, mostly because that sounds like something very specific to you, Michael. I think there’s a general point to be made about GP’s bringing the weirdos out of the woodwork and to be ready for literally anything, but Lich? I think this may just be you taking one for the team, and I thank you for that (I don’t want to play against Lich or Pillow Fort round one).

I disagree that you should always have hate for .06% of the format. This is not Extended. Every format follows its own sideboarding rules.

You distinctly ignored my rule 4, aka the Lich rule. Why does everyone ignore the Lich rule?

Michael Warme has it right. Graveyard hate is good against a lot of decks, and Dredge is increasing in popularity online and in paper, so while you should have had it before to combat Grixis, Living End, and Abzan Company now you really need it or you risk losing to Dredge.

The rule came about because back when Extended existed every few weeks Dredge would suddenly win a lot of tournaments because players hated it out then got complacent and took out their hate. I made sure to always have hate so that when that surge inevitably happened I was ready. While Dredge decks have always been around in Modern (Aggro Loam has been a frequent ‘surprise’ deck in the Colorado metagame for years now) we’re seeing a dramatic increase since Shadows came out, so once again it’s essential to have the means to beat that deck.

Among other things, yes. When the Beginners Guide version doesn’t cover everything I’ll follow up the week after with the more advanced tips and tricks.

The point he was making is that you should pack some kind of graveyard interaction–it’s a bullet against some decks, and useful against most to some extent or another.

My personal rules for sideboarding in ALL nonrotating formats:
1. There is always a “dredge deck” of some kind in the format and you should have hate. Sometimes it plays Bazaar, sometimes It plays cranial plating, sometimes it’s goblin charbelcher. There’s a linear cannon deck that operates on a different axis from everybody else, and you can play hate for it, and you can win the matchup. Don’t leave the hate and expect other people to hate out the cheese deck.

2. Play graveyard, artifact, and enchantment interaction. This format has a large card pool, and your opponent could be doing all sorts of noninteractive shenanigan things. Most tricksy things involve using artifacts, enchantments, or the graveyard, so having SOME amount of interaction with each of these zones/permanent types is important.

3. Understand what the disruption is in the format (thoughtseize? Force of Will? Mental Mis-Step?) and have sideboard plans that line up favorably against these interaction points (flashback spells/cavern of souls/2 drop creatures).

4a. If Lich is legal in the format, some bonzo in the GP is playing it in their mainboard, and you will be paired against them round 1. I might or might not be 3 for 3 on this statement being true at legacy GP’s.

4b. If Lich is not legal in the format, that bonzo is playing modern enchantment prison, you’ll get paired against them round one, and they will curve out in such a way that you question why this deck isn’t tier one because it clearly has every possible angle covered.

Yeah, I was a little curious about that too. Maybe the author’s LGS has lots of dredge for some reason?

The reason you give for not excluding Dredge hate from the sideboard:
“It’s easy to fit a few Relic of Progenitus into any deck, so just make sure that you aren’t cold to this deck.”

That’s just not true. A lot of decks DO struggle to fit a pair of Relics in their sideboards, which is why we don’t see Relic in that many sideboards. If it was so easy, more players would do it. But it’s not easy — instead, it costs precious slots that could be dedicated to decks that hold more than a single metagame percentage point, as opposed to ones that hold less.

This may not be a good “endpoint”, but it’s certainly been helpful for me to get started.