I've never done arbitration, but I did pursue legal action against a contractor that defrauded me and I think the logic is basically the same.
Lawyers are constantly evaluating the cost (to their client) of further pursuing a claim vs the likelihood of succeeding. They know how much it will cost in time if it goes all the way to litigation/arbitration, and they have a fair idea what the outcome will be, but it's uncertain. They are always going to try to settle long before the end-game unless they are trying to make a point and it's worth wasting a bunch of money on.
In my case, my lawyer was very convinced that they were likely to settle for an amount basically up to about 75% of the amount it would cost to litigate. My claim was solid (he DID defraud me) but there was some possibility that a trial would find otherwise, so it was better for them to settle up to the point that it became worthwhile to roll the dice.
The emails between their lawyer and mine were about details of the dispute, not about the financial calculation I outline above, but that was just window dressing; the real decisions were made based on how much it would cost to litigate vs how much they would pay to settle.
The only place the facts of the fraud even came into it were as part of the calculation of how likely we were to prevail in court, but it was not at all precise; more like "given the facts, we are pretty likely to win but not certain".