In sources, such as Chapin’s Next Level Deckbuilding, you can find easy rules of thumb for how you should count things like mana-elves, cantrips, and ramp spells in your deck. Generally those rules are simple, such as they should be counted as a half or a quarter of a land.
Well, since fetches are going to be in Standard again, lets try this out.
So, I wrote a small program to try this out. The list used is this, although I’ve simplified the list for easier analysis:
[cardlist]
[creatures]
4 Tormented Hero
4 Gnarled Scarhide
4 Bloodsoaked Champion
4 Mardu Skullhunter
4 Pain Seer
4 Spiteful Returned
2 Mogis’s Marauder
[/creatures]
[spells]
4 Thoughtseize
4 Despise
4 Crippling Blight
[/spells]
[lands version 1]
4 Bloodstained Mire
4 Polluted Delta
14 Swamp
[/lands version 1]
[lands version 2]
22 Swamp
[/lands version 2]
[/cardlist]
The test we are trying out is simple: how many lands do we draw in a set amount of rounds (10).
The program takes a mulligan whenever there’s zero lands or four or more lands. We assume that you always use the fetches first, even if this is not necessarily the case in real life. I just couldn’t be bothered to program in the rules for figuring out the situations where you don’t want to do this.
Here’s the number of lands on each of your first ten turns (you’re on play) with version 1 lands:
0.99992
1.8463
2.48568
2.90142
3.22642
3.52124
3.80418
4.08784
4.36606
4.65026
On your first turn, you have less than one, because sometimes you mulligan to oblivion and can’t play a land. For comparison, version 2:
0.99992
1.85994
2.56562
3.1168
3.56926
3.97866
4.37242
4.7519
5.12534
5.49802
In short, by round 10, you’ve drawn .85 spells more than without the fetches. Sure, you’ve taken some damage in the mean time, but with a deck like this, that’s not a real cost in most cases.
How about if we put in a couple of more [scryfall]Mogis’s Marauder[/scryfall]s and take out two swamps?
0.9998
1.81302
2.39228
2.75688
3.03454
3.2894
3.5351
3.78124
4.02654
4.27516
and
0.99976
1.82862
2.48004
2.97222
3.3698
3.73288
4.0833
4.42164
4.7583
5.09598
Seems to me like 20 lands are enough for this deck (and I don’t think anyone would really disagree). Again, the number of lands drawn by round ten is about the same.
One thing you should note: In the end, this isn’t about one game in particular, but rather series of games. This is about hedging your chances. In tournament play this might mean a win here and there, but that’s what you need to make top 8 of a major tournament: finding the best possible percentages.
Whether you find the loss of the second land in certain marginal number of cases too much depends on you.
Here’s the code I used. Use it freely. I had to zip it, because WordPress doesn’t like uploading .php-files, which is pretty understandable.
(And yes, sure, some other languages would support this sort of stuff better, but I have everything ready for PHP, so that’s what I use for this sort of small projects.)