`max_errors` clarification - its use and purpose

I want to clarify the purpose of max_errors I understand that it is an option to specify the number of erroneous terms in a given phrase. In my use case, I would always want all terms to be corrected so for example:

paracetmol tummour CLODRoONATE

If I specify max_errors as 2 I get the following output:

"suggest": {
"phrase": [
{
"text": "paracetmol tummour CLODRoONATE",
"offset": 0,
"length": 30,
"options": [
{
"text": "paracetamol tumour clodroonate",
"highlighted": "paracetamol tumour clodroonate",
"score": 0.0000011286303
},
{
"text": "paracetmol tumour clodronate",
"highlighted": "paracetmol tumour clodronate",
"score": 3.56549e-7
},
{
"text": "paracetmol tumour clodroonate",
"highlighted": "paracetmol tumour clodroonate",
"score": 1.8145049e-7
},
{
"text": "paracetamol tummour clodronate",
"highlighted": "paracetamol tummour clodronate",
"score": 1.7691919e-7
},
{
"text": "paracetamol tummour clodroonate",
"highlighted": "paracetamol tummour clodroonate",
"score": 9.003551e-8
}
]
}
]
}
}
However I would always want as many terms to be corrected as possible so in this instance I would specify 3, since a query string can be any number in length, I'm planning on setting this to a large number such as 200 to ensure as many terms are corrected as possible. So the question is what exactly is the purpose of max_errors why would you want to set a limit on it? Perhaps I'm misunderstanding its purpose of what it is intended for.

If it is set to a fraction (ie >0 and <1) rather than a number >=1 then it acts a percentage of the provided terms.

The docs [1] warn of a performance hit but I presume it might also mean that you have suggestions that aren't remotely close to what the user typed.

[1] Phrase Suggester | Elasticsearch Guide [2.3] | Elastic