Donk Bet Raising Success Rate...

[de-coder]

You know you're really starting to piss me off now. I've basically just been ignoring your blather about this because that's the way I roll generally, but I've been stewing over this (and your other idiot comment). You don't find this useful - fine - shut the fuck up. I've poured hours (too many probably) into this to try and help figure this out (not the sql stuff, but the poker stuff - I've had to figure the sql stuff out, but I'm doing it for the poker...) because some of us are curious to see how this works.

If you have something constructive to say, great. You want to help get us better information, great. You want to keep on trying to invalidate the work we're doing - not great.

[Talking Poker]

+Rep * {Infinity}

[Wes]

Really easy, although you wouldn't have the l33tness because you don't know have to know how to do sql queries.

[Talking Poker]

I updated the SQL in the first posts with "x.real_player_id" instead of "x.player_id."

So here are my personal results with raising donk bets. I'm using my Troll Pie ID only, so these should only be hands from 1/2 NL - 5/10 NL (all 6 max) games on Full Tilt since last October or so.

I raised 362 donk bets.
Donk has folded 189 times.

So, that's 52% of the time that I've taken down the pot right there.

I know one person will find this useless, but I think this is awesome, since in my head, I was guessing this number (for me, obviously) to be slightly over 50%.

For the stupid among you, please be aware that I do not think I have proved always raising a donk bet to be profitable or anything else like that. But I have proved that for me, when I have choosen to raise a donk bet, it has indeed been +EV... at least as far as the bet itself is concerned and ignoring the rest of the hand (where I am almost certain I have won a lot more money that I have lost).

I'd be interested in seeing more people's numbers, especially the people who speculated what their results were in the other thread.

[Wes]

OK, you proved two people wrong in their assumption of just general donk betting and how often a person folds doing it. Now, can we please expand upon this into some more thorough analysis that will help people more?

Can you filter your actual winnings when you do this play? It might be a little more useful for deciding if just auto raising donk bets is a better way to approaching the hand than say evaulating what the donk bet means and acting accordingly.

And then, when you are done with that, filter on how you are doing when you raise a donk bet from a person with a vpip over/under 30 and the profitably of that play.

[de-coder]

to find the level you want try this

select * from game_level

that will give you all the levels you have in the DB and their game_level_id

then use

select x.game_id from game_players x join game_players y on x.game_id = y.game_id and y.fbet = 1 and y.praise != 1 and y.ffold = 1 join game z on x.game_id = z.game_id and z.game_level_id = game_level_id here where x.raised_first_pf = 1 and x.fraise = 1 and x.fcheck != 1 and x.real_player_id = player_id here

to those who actually appreciate my work - you're very welcome. This is such a fascinating topic. There's a lot of good stuff in there.

Here's a question:

how often is a limp - re-raise AA?

To be fair I'm not talking about pure limps, but people that call the current bet, get raised at some point behind them, and then re-reraise pre-flop.

Any guesses?

[Talking Poker]

You have way more data that I do. By all means, go for it.

[Talking Poker]

Stats are fun.

This is a tricky one though, especially with the broad definition (maybe this should be a new thread - but I guess there's not too much to discuss with this one). I think the number is pretty high... how about 42%?

[lightfungus]

I'm doing a decent amount of SQL at work right now and would love to help out with this but I don't have the time right now. So I appreciate it de-coder. And work also blocks this site

[de-coder]

Keep in mind I'm not actually a sql guy - I'm just kindof flying by the seat of my pants and reading a lot of sql docs (I also asked a guy at work to point me in the right direction). If you have any suggestions as to how to do things better, or if you see any problems with it let me know.

There's more stuff in there too if we have any ideas for things we'd like to see (see above for 1 fun stat...).

[Wes]

Use it.

[Reel Deal]

Finally getting around to reading this after my Cali trip... thanks TP and de-coder, this is very interesting stuff.

And BL, you really, really need to learn how to deal with people w/o being a complete ass... trust me, it will help you later in life when you eventually crawl out from the bowels of your parent's basement.

[de-coder]

No other guesses on this?

TP guessed 42% - anyone else?

[Reel Deal]

62%

[meanness]

Good thread, thanks for the work.

I don't have any data, but I'm sure my 5% guess was strongly influenced by the way I was running at the time, and not truly representative of the general trend. Thanks to this thread, I've been experimenting a lot more with raising the donk bettor.

On a drawy board, I'm starting to find more success - not with the initial fold, at $100NL anyway - but if the turn blanks and I fire again, I'm winning the pot consistently.

[Talking Poker]

Exactly.

Nice to know that not everyone finds this thread useless.