When is it better to underpromote?

....when u can't overpromote ??

Hmm.  What about a situation where my the promotion square is attacked by my opponent and by me, say by two rooks.

If I promote to a queen, my opponent will of couse take it with the rook, and I will in turn take their rook.  I am up a rook.

OTOH, if I promote to a bishop or knight, my opponent may NOT want to trade their rook for it.   Thus the endgame is rook+bishop vs rook instead of rook vs. nothing.     

From my POV, would I ever want the net effect to be two pieces vs one opponent, instead of one piece vs none?

If so, that's another reason to under-promote.

Probably not, but in any case, why let them make the choice?  By promoting to a queen, you (more or less) force them to take.  If you underpromote, they still have the option to take, if that's the better move for them, as well as the further option not to, if that's better (which most times, it probably is).  It sounds like you're hoping they'll make a choice against their interests -- to not take in a position where it would be better if they did.  That's called "hope chess," and it's not recommended.

Granted, it leaves a decision up to my opponent.  However, it's not quite hoping they do a bad thing for them.  It's offering a calculated trade which may be rejected.

If I can determine that a rejected trade is better for me it's not hope chess,   I could underpromote and benefit more than I would otherwise.   Outcomes:

1. Lose the un-promoted pawn.  Bad for me.

2. Promote to queen, trade queen for rook.   Net loss to opponent of a R

3a. Promote to N/B/R.  Opponent takes.  Net loss to opponent of a R.

3b. Promote to N/B/R.  Opponent rejectes.  Net gain to me of N, B, or R.   

If the added flexibility of 3b is better in a particular circumstance....?

But it's not flexibility -- except for your opponent.  Flexibility is when you have the choice.  You don't get to make the choice between 3a and 3b.

'Choosing' to play a Rook + Bishop vs Rook ending instead of a Rook vs nothing ending doesn't seem like the best choice out there!

Oh, yeah I understand that it's not a good general principle.

This thread started with a list of rare or contrived reasons to underpromote.   Going from 1:1 to 2:1 instead of 1:0 seems like it might have application, even if only in a puzzle.

If you want to see all the underpromotion uses in one puzzle twice, google for the Babson task. I guarantee it will never happen to you in an actual game!

96.9%Q，1.8%N，1.1%R，0.2%B

"Promote to a knight to queen with check ..."

Uhm, if you promote to a Knight, you are not "Queening" now, are you?

You may be promoting, or possibly "Knighting" a Pawn, but you are not queening if you are not getting a Queen!

If my opponent is threatening checkmate, then I will in some positions under promote to a knight if it means that I can win material with check to prevent my opponent from checkmating me when promoting to a queen would lead to my getting checkmated.  Also in FFA 4 player games I may promote my pawn to a knight if it is doomed to get captured to minimize the amount of points the opponent who captures it will get from capturing it.

So would this considered an overpromotion ? lol !!

1. c8 (Q)

okay  ....now. I wantchu to........oh nvm.

underpromotion is only done when you are low on time and trying to avoid checkmate

or when there's a variation that gives checkmate

or you're trolling your opponetn

when you underpromote to a knight sometimes you can fork pieces

If u overpromote and oops a stahayleemate u shouldnt feel ashamed. dumb maybe but not ashamed.

I would imagine most players who play regular Blitz chess have auto-Queen selected so being low on time is unlikely to be a reason for under-promoting if you have to go into your settings and switch off auto-Queen before making your move!

to be forced to promote a pawn to a piece feels like authority is besetting ideal b/s upon me.

I dont like it and i never will. I wanna go back to across the pond non-antebellum 1860 where billy steinitz fought galantly (sadly in vain) to maintain & sustain that a pawn can remain a pawn if it canwell pleezed.

That said, less than 3% of the time we u/p. now. Id like a game study where pushing a pawn and leaving it one as such can be a nicety. 

ur hereby challenged....all u composers of the beneviolent arts. i mean benevolent. i think.

This is an example of a situation, in which promoting to a fairy piece would be better than promoting to a standard chess piece.  Promoting to any standard chess piece would loose to the checkmate, but there is a particular type of fairy piece that, if white could promote to would be winning.

no? Black has a mate in 11