Bolty's Note: While statistics and chances to find set, unique, rare, magical, and normal items had been known for a while when wearing +magic find% gear, studies had generally been done in 50% degrees. For example, the chances to find items would be calculated for wearing +100%, +150%, and +200% magic find, etc. But towofu would make a discovery that surprised everyone - that a mere 1% difference in magic find chances can make a huge difference.

I experimented about +Magic find%, and noticed there is a big difference between +150% and +151%.
Please see the following; these are the results of how many magical items were dropped.
I picked up 100 exceptional items, except throwing weapons that can not be magical. (Magical means Unique, Rare, Set and Magic. non-Magical means superior.)

Bolty's Note: Throwing weapons could not be magical in standard Diablo II, when this discovery was made.

OPEN-SINGLE game in Hell Arcane Sanctuary (mlvl=69)
+150% MagicFind: got 65 magical items and 35 non-magical.
+151% MagicFind: got 100 magical items and 0 non-magical.

Tested on realm also.

US-WEST REALM game in Hell Arcane Sanctuary (mlvl=69)
+142% MagicFind: got 61 magical items and 39 non-magical.
+156% MagicFind: got 100 magical items and 0 non-magical.

The following formula (by Jarulf) says there is no big difference between +150% and +151%.
Magic: 1/X, X = (3-mlvl/100)*100/+GettingMagic

Bolty's Note: Actually, this formula wasn't Jarulf's, but it is generally correct.

I had more experiments, and picked up about 1000 exceptional items:

OPEN-SINGLE game in Hell Arcane Sanctuary (mlvl=69)
+150% MagicFind: 6 Unique, 46 Rare, 48 Set, 619 Magic, 281 Superior
+151% MagicFind: 4 Unique, 43 Rare, 54 Set, 905 Magic, 0 Superior

Don't you think something has occurred between +150% and +151%?
Could you give me some information about this, please?

Bolty's Note: Jarulf responded with a little information on why this could be the case:

OK, the check for being magic or not (if it fails, it goes on to check for superior and so on) is, for exceptionals, done with a base chance of one in 3 for being magic. Meaning that of those that do not become unique, rare, or set, 1 in 3 would be magic. The level of the monster does NOT matter, since its level is divided by 100 first (and rounded down, thus it always ends up as 0 and wouldn't affect it). Now, we ONLY handle the +151%/+150% first, so skip the base chance. The way the game calculates the chance for the modified value is:

BASE*100/MF

Now, base is 3 as I said. In addition, we round down after the division, so we get:

+150% MF: 3*100/150 = 2, which means a 1 in 2 chance.

+151% MF: 3*100/151 = 1.987 which rounds down to 1, or a 1 in 1 chance.

There is the explanation.

Bolty's Note: Spirea added even more to the discovery:

I did some research on chances for exceptional rares. At 151% MF, we see a significant drop in the chance of getting exceptional rares (about 20% less compared to 150% MF). So, the question then is, should we stop at 150% MF?

Mlvl 90 <=> 188%
Mlvl 70-89 <=> 194%
Mlvl 50-69 <=> 196%
Mlvl 26-49 <=> 198%

Above are the break even points for the various monster levels. So, if you can get MF to 198% or higher, your chance of finding exceptional rares will be better than 150% MF. If you stay in act 2 hell or above, 194% is sufficient. If you cannot meet the break even point, it is better in terms of finding exceptional rares to just stay at 150%.

Other factors that might make you want to go to that 151%, even though the chance of exceptional rares may be lower:
1. You might just want magical items.
2. You might want rares that aren't exceptional (especially rings and amulets).
3. You can't find items to add up to exactly 150%. In which case, you should calculate to see what is the best combo.

Bolty's Note: Magic Find now works differently in Diablo II v1.08 and above, but the discovery made back then was correct for its time.