.. creating a holdem odds software .. need advice/critisism ..

[denarced]

so
writing this one with C ( as ANSI as possible )
need advice/tips/criticism
hopefully all the above
don't hold back, I'll read em all

the idea is that one day there will be ui/gui
user will choose:
table size
what hole cards he has
and
possibly
flop
turn
river

the software will then make about 10'000 rounds
and then give a percentage of wins

also,
pay attention to the random number generator
time-function will give a new srand-seed only once per second
not sure if this is random enough

the code:

#include <stdio.h>
#include <stdlib.h>
#include <time.h>

typedef struct kortti /* a card */
{
	int maa; /* suit 0-3 */
	int arvo; /* value 0-12 */
} KORTTI;

typedef struct kasi /* a holdem hand(2 cards) .. not yet in use */
{
	KORTTI a;
	KORTTI b;
} KASI;

typedef struct ll /* linklist for the deck of cards */
{
	KORTTI item;
	struct ll *seuraava;
} LL;

LL * luo_ll(void); /* create linklist eg the deck */
KORTTI * arvokortti(LL **linklist); /* pick a random card from from created by luo_ll()*/
int tuhoakortti(KORTTI *kortan,LL **linklist); /* destroys a card from the deck */
void virhe(int status); /* for serious failures and program shutdown */
void tulosta_kortti(int mode,int maa,int arvo); /* given mode(0 or 1), suit and value, will print the card in an understandable way */
float calc_odds(KORTTI *kortit,int k); /*args: player1's cards,[flop,turn,river] */

int main(void)
{
	int k = 3;
	KORTTI *kortit = malloc(k*sizeof(KORTTI));
	(*kortit).maa=0;
	(*kortit).arvo=0;
	(*(kortit+1)).maa=1;
	(*(kortit+1)).arvo=0;
	(*(kortit+2)).maa = 2;
	(*(kortit+2)).arvo = 0;
	float res = calc_odds(kortit,k);
	printf("float res = %f\n",res);
	return EXIT_SUCCESS;
}
float calc_odds(KORTTI *kortit,int k)
{
	
	srand(time(NULL));
	LL *ptr = 0;
	ptr = luo_ll();
	if(ptr == 0)
		virhe(2);
	int i,t;
	for(i=0;i<k;++i)
	{
		t = tuhoakortti((kortit+i),&ptr);
		printf("destroia tulos: %d\n",t);
	}
	/* yllä olevan tuhoamisen voisi yhdistää korttien siirtoon pakasta */
	/* luotavaan tietorakenteeseen eli kortit siis "siirretään" */
	/* luo tietorakenne pelaajien käsille ja flop/turn/river */
	/* jaa kortit muille pelaajilla ja flop/turn/river */
	ptr = 0;
	return 3.14;
}
int tuhoakortti(KORTTI *kortan,LL **linklist)
{
	int i;
	LL *edellinen = *linklist;
	*linklist = (**linklist).seuraava;
	for(i=0;;++i)
	{
		if( (**linklist).item.maa == (*kortan).maa && (**linklist).item.arvo == (*kortan).arvo )
			break;
		else
		{
			edellinen = *linklist;
			*linklist = (**linklist).seuraava;
		}
	}
	(*edellinen).seuraava = (**linklist).seuraava;
	free(*linklist);
	return EXIT_SUCCESS;
}
KORTTI * arvokortti(LL **linklist)
{
	int rint = ( rand() % 1000 ) + 100;
	int i;
	LL *edellinen;
	KORTTI *newcard = malloc(sizeof(KORTTI));
	if(newcard==NULL)
		virhe(59);
	for(i=0;i<rint;++i)
	{
		edellinen = *linklist;
		*linklist = (**linklist).seuraava;
	}
	(*newcard).maa = (**linklist).item.maa;
	(*newcard).arvo = (**linklist).item.arvo;
	(*edellinen).seuraava = (**linklist).seuraava;
	free(*linklist);
	return newcard;
}

LL * luo_ll(void)
{
	LL *eka = malloc(sizeof(LL));
	if(eka == NULL)
		virhe(59);
	LL *current = NULL;
	current = eka;
	(*eka).item.maa = 0;
	(*eka).item.arvo = 0;
	int i;
	int maa=0;
	int arvo=1;
	for(i=0;i<51;++i)
	{
		LL *new = malloc(sizeof(LL));
		if(new==NULL)
			virhe(59);
		(*new).item.maa = maa;
		(*new).item.arvo = arvo;
		current->seuraava = new;
		current = new;
		if(arvo<12)
			++arvo;
		else
		{
			arvo=0;
			maa+=1;
		}
	}
	(*current).seuraava = eka;
	return eka;
}
void virhe(int status)
{
	if(status==59)
		printf("malloc problem\n");
	if(status==2)
		printf("linklist creation failure\n");
	exit(1);
}
void tulosta_kortti(int mode,int maa,int arvo) /* mode 0 = short, 1=long */
{
	if(arvo>0&&arvo<10)
	{
		printf("%2d",arvo+1);
	}
	switch(arvo)
	{
		case 0:
			printf(" A");
			break;
		case 10:
			printf(" J");
			break;
		case 11:
			printf(" Q");
			break;
		case 12:
			printf(" K");
			break;
	}
	switch(maa)
	{
		case 0:
			if(mode==0)
				printf("H");
			else
				printf(" of Hearts");
			break;
		case 1:
			if(mode==0)
				printf("S");
			else
				printf(" of Spades");
			break;
		case 2:
			if(mode==0)
				printf("D");
			else
				printf(" of Diamonds");
			break;
		case 3:
			if(mode==0)
				printf("C");
			else
				printf(" of Clubs");
			break;
	}
}

[WingedPanther]

I have an issue with the basic strategy you are taking in your program. You would do better to understand the probabilities and expected values involved, and code against those than a random sample of 10,000 hands. In a game with 4 players there are 733,055,400 possible ways for the other 3 players to have their cards on the final round of betting. That number is higher on earlier rounds. your 10,000 random samples are not statistically significant.

[denarced]

I have an issue with the basic strategy you are taking in your program. You would do better to understand the probabilities and expected values involved, and code against those than a random sample of 10,000 hands. In a game with 4 players there are 733,055,400 possible ways for the other 3 players to have their cards on the final round of betting. That number is higher on earlier rounds. your 10,000 random samples are not statistically significant.

hmm .. suppose I could do a few tests and see which number would be right .. I thought 10'000 would give me a reasonable accuracy .. somewhere like +-1% error .. which is enough .. I suppose I could do calculations and increase the number of rounds until the difference of odds are not meaninful anymore ..

[WingedPanther]

The above was for when you have the river card, it gets worse the earlier in the process you get. To have a valid test, you would need to implement several multiple of 1,136,464,726,320 (the number when 3 cards are showing in the middle). You're looking at a minimum of 100 trillion trials to get meaningful test results.

[denarced]

The above was for when you have the river card, it gets worse the earlier in the process you get. To have a valid test, you would need to implement several multiple of 1,136,464,726,320 (the number when 3 cards are showing in the middle). You're looking at a minimum of 100 trillion trials to get meaningful test results.

hmm .. I have to disagree .. surely I don't have to know each and every possibility to know the odds ?? ..

[WingedPanther]

You're correct, but to estimate the odds through trials requires performing enough trials to estimate the true odds with confidence. Since certain hands have a VERY low occurrence, relative to the number of possible hands, you will need at least as many trials as the actual number of possible results.

For comparison: There are 40 ways to get a straight flush on a 5-card poker hand. There are 2,598,960 possible 5-card poker hands (.15 in 10000 probability). On 10,000 trials, the probability of getting any straight flushes is 14%. That means there's an 86% chance that you would conclude there is 0 chance of a straight flush and a 14% chance you will conclude there is a 1 in 10000 or higher probability of getting a straight flush. Either way, the conclusion would be wrong. Note: I am producing these numbers with the computing power of a $10 calculator, so the calculations involved aren't overly arduous.

[denarced]

hmm
It would be interesting to see how much power there is in computers these days ..
I wonder how much time would it take to do all the trillions of possibilities
Processors are pretty fast these days so I imagine not as long you'd think

[denarced]

But I do see your point
There is a better way to approach probabilities in poker than to do random trials
I just don't know what it is

I might do it this way just because I'm a beginner and this way is easy enough for me

[WingedPanther]

Realize, too, that you're talking to a guy that taught probability/statistics

[denarced]

Excellent
So then you know how to approach something like this .. how ?
( if it's not too complicated )