.. Every time you reset the withdrawal rate you decrease the probability of success...

It is more complicated and it isn't that cut and dry.

The problem is that if you reset it after the first year, then instead of looking at a 30 year window, you are now looking at a 29 year window that the withdrawals will be made over. The shorter required time frame would tend to offset the higher withdrawal rate. Since the time frame has been decreased by a year the original 30 year model does not apply any more and can't be used to predict the probabilities of success and a new 29 year model would need to be found to estimate the odds of success.

You really need to also look at the actual dollar amount involved to see if it is significant.

Assuming inflation is 3%, the inflation adjusted withdraw would be $82400.
Recalculating 4% of $2.1 million is $84000.

The difference the two is $1,600, which is 0.076% of the $2.1 million and within 2% of the lower inflation adjusted amount, which could easily be more than offset by the year shorter withdrawal period. The models used to calculate the SWR aren't nearly precise enough to make this a clearly bad choice.

For a single year I really don't see any problem with upping the withdraw a bit, but this should be closely watched at say 5 and ten years to make sure that it is still reasonable then.

Greg