Is there a recommended amount of land worth keeping in your deck? I've read between 20-25.
26 isn't outlandish, there are decks that like to go as high as 28.
These are the two go-to articles when it comes to the math involved:
https://strategy.channelfireball.co...you-need-to-consistently-hit-your-land-drops/
https://strategy.channelfireball.co...do-you-need-to-consistently-cast-your-spells/
Basically, you're trying to hit a sweet spot which balances three things for your deck - getting consistent land drops, getting the right colors of mana, not getting too much mana flood.
Getting consistent mana drops means being able to play a land on turn 1, turn 2, turn three, etc. Some decks can run optimally at just two lands, some want to stop at three-four, some decks
want all the fucking lands (typically stuff that ramps by playing extra lands). Then, you weight it against the chance of mana flooding.
There's a handy table in the first article, for example you want to build an aggro deck and want to decide between 20 and 22 lands:
20 lands 20 [98.9% / 98.3%] [86.4% / 79.6%] [65.8% / 55.2%] [44.1% / 33.7%] Flood: 3.9%
22 lands 22 [99.5% / 99.2%] [90.5% / 84.7%] [73.9% / 63.7%] [54.3% / 43.0%] Flood: 7.3%
As you can see, at 20 lands you have really good odds of getting three lands out (sometimes you'll wait for it for an extra turn or two), and then it falls of. If you might as well never see the fourth land, or can help dig for it with scry or draw, it's a good number.
You also get a very low chance to be flooded (though Arena shuffler will happily oblige)
In comparison, 22 lands gives you a 10% increase in consistency in drawing land number four. If your aggro deck is mana hungry and needs to get those four mana out, say you will want to reliably escape Phoenix of Ash,
may be worth considering going 22, even though the flood chance is almost doubled, it still looks manageable.
Can you go lower with aggro? Some play 18-19 lands. Reducing the risk of mana flood, thus you should pretty much topdeck nothing but gas. The trade off is you will get bad hands from time to time, forcing you to mulligan.
There are decks that can pull this off, e.g. mono blue tempo with plenty of draw from curious obsession can reliably get three lands this way if it gets its engines going. Most of its spells also cost only 1-2 mana,
so it can do anything it needs to with just two lands out.
Another example was a mono red deck with Wayward Guide-Beast which lets you bounce a land back, replay it untapped and get mana again. Fun deck to try in historic, it pretty much only
has 1 mana stuff. Lets you go really low on lands.
Another comparison, 24 and 25 lands:
24 lands [99.8% / 99.6%] [93.5% / 88.7%] [80.6% / 71.3%] [63.8% / 52.2%] Flood: 12.2%
25 lands [99.8% / 99.7%] [94.6% / 90.4%] [83.5% / 74.7%] [68.2% / 56.7%] Flood: 15.2%
As you can see, the % chance for dropping lands turn 3-5 are pretty much the same at this point, yet you end up with a rather high hance to flood (15%).
Personally, I don't like running 25 lands, I tend to just go 24 if it's important to go to 4 lands, or 26 if I'm playing control.
MtGA tends to reflect that math, there used to be some talk about exploiting the way shuffler fudges the draws in Bo1, but you should just focus on giving yourself enough mana
instead of trying to game the system and seeing how low you can go. This way you only need to worry about mana flood (when it happens, it happens, give yourself some draw, scry or mana sinks - cards that can use lots of mana), instead of both flood and short.
This is where the second factor comes in - giving yourself enough colored mana sources. At 20 lands, you might need to use all the dual lands you can get your hands on to get good odds of getting two colors (it'll be super tight, even then
it might be difficult to achieve double mana costs like 1BB and GG). Three colors is pretty much a no-go because there isn't enough lands for it. You need 22+ lands here.