From my experience more times than not that less than +3 is irrelevant win or lose.
I assume you mean football
In case of hockey, most frequent outcome is the favorite beating the underdog by exactly 1 goal. Obviously there are a very wide range of outcomes, but that is the #1 most frequent outcome.
An event with a 38% chance of happening is gonna hit about 38% of the time, just like an event with a 70% chance of happening is going to hit about 70% of the time over the long haul. The question is if you can consistently get a price that pays out better than the true probability of something happening.
Like if an event's price is +190, but the true price is more like +175 based on the actual probability of the event, that's when you're in business . Obviously identifying those opportunities is the hard part
People are gonna play a +170 ML over a -130 +1.5 because hitting the former is more fun, pays out bigger, but reality is if you flat bet each of those 1000 times you're coming away with about the same result in terms of profitability