Advanced search

PFR Analysis

Discuss PokerTracker, GT+, PokerAce and Others

Moderators: excession, LPF Police Department

PFR Analysis

Postby APerfect10 » Tue Apr 26, 2005 2:26 pm

I mentioned a week or two ago that I wanted to determine the optimum amount of times to pre flop raise. With the adjusted auto rate rules that Excession has put together and I have helped critique ratings are no longer based on PFR% and I personally think that is fine although I read a complaint. Combine these two things and I really wanted to analyze pre flop raises to the fullest.

I took the two most profitable types of players, TA (VP$P <22%, Flop Agg 1.8+) and sLA (VP$P 22-35%, Flop Agg 1.8+) to analyze their PFR%'s and corresponding PTBB/100 rates. I decided to use these two player types for two reasons.

1. They are the most profitable
2. Recently I am finding myself borderline between the two (so they apply to myself the greatest)

For TA, I broke them down into 10 categories of increments of 1 with a cap of 10. I would have liked to take this further, possibly to 15% PFR; however, their just were not enough TA who raised that frequently so the sample size was too small to be significant therefore I lumped everyone who raises more than 10% into the same cateogry. For sLA, I broke them down into 15 cats of increments of 1 with a cap of 15.

Notes regarding the graphs below:

1. The Y axis is PTBB/100 winning rates.
2. The X axis is their PFR%. 5 corresponds to PFR%'s in the 5.01-6.00 range.
3. For TA, 10 = 10+ PFR% and for sLA, 15 = 15+ PFR%
4. The lines (arcs) are best fit, polynomial trend lines.

Image

For TA players, it appears that the optimum range to raise is somewhere in the 5.5 - 7.5% range. Below 5.5% and above 7.5% winning rates are not at their peak. 6.50% seems to be peak with a PTBB/100 of about 5.25.

Image

For sLA the graph is similiar, however, the peaks are greater than that of a TA meaning they are making more. Another interesting thing to note is that the peak percentages are shifted over by a few percent in comparison to the TA. The optimum range is 6.5 - 9.5% with the peak occuring at 8% with a PTBB/100 of about 7.5.

Conclusions:

Although the data was not as clean as I had hoped, I think that the polynomial trend lines give us as good of a prediction as it gets. TA players optimum range is 5.5-7.5% with the peak occuring at 6.5% with a PTBB/100 of 5.25. The sLA wins more and should be raising a little more frequently. The sLA optimum range is 6.5-9.5% with the peak occuring at 8.0% with a PTBB/100 of 7.5-8.

It is clear why the sLA is winning at a higher rate since they are playing more hands that have a +EV value; however, I am unsure why the optimum %'s are shifted higher for the slightly loose aggressive. Could it be due to the fact that sLA raise more with lower suited connectors while the TA is raising more with paint cards? I'm not sure although that is the first thought that comes to mind.

What is the optimum amount of times that you should raise pre flop? I personally would aim for around 7-8% regardless of your player type; however, make sure that you are picking your spots and play smart post flop. Post flop aggression is more important than pre flop aggression. As long as you are somewhere in the optimum range you should be maximizing your profits.

How does this apply to the new auto rate rules? This just proves the point that the previous rules of dividing up the player types by pre flop aggression with an aggressive/passive mark at 7% was purely meaningless. The only reason that the previous pre flop aggressives tended to win more than the labled passives is due to the fact that the <1% PFR players were pulling the passives down further than the extreme aggressives.

To get an accurate auto rate rule by using pre flop raises, you would need to divide them into two groups. Goodbye to the passive and aggressives pre flop. Instead you should have one group that is in the optimum range and one that is on the outer edges. This would provide the most meaningful results. Without digging much deeper into this, they would be as follows:

Optimum: 5.5% - 9.5%
sub-optimum: 0-5.49% and 9.51%+

Maybe once Pokertracker Pat gives us a few more player types we can incoporate PFR%'s back into the auto rate rules using these optimums. Until then, the new rules are the best that we can do with so few player types.
User avatar
APerfect10
 
Posts: 208
Joined: Mon Feb 14, 2005 2:11 pm
Location: Pennsylvania

Postby gdaviet » Tue Apr 26, 2005 5:56 pm

That analysis is awesome. I'm a numbers geek, and that impresses me. What limits is your data from?

However, I still like my auto-rate rules with the PFR split. I'm don't split at 7% though.

Again, fantastic work. WOW!
User avatar
gdaviet
 
Posts: 61
Joined: Thu Apr 14, 2005 7:39 pm

Postby APerfect10 » Wed Apr 27, 2005 2:37 pm

Thanks for the kind words.

My database is 1GB+ with mostly $50 NL and some $100 NL mixed in. I dont suspect that these numbers should change all that much, if any, at other levels of NL. PL and Limit may be an entirely different beast though...
User avatar
APerfect10
 
Posts: 208
Joined: Mon Feb 14, 2005 2:11 pm
Location: Pennsylvania

Postby k3nt » Fri Apr 29, 2005 11:42 am

On TA vs sLA preflop raises. If you take preflop raise and divide it, not by total hands played, but by flops seen, you may get very similar numbers.

For instance, a TA with PFR 5% and VP$IP 20%: 5% / 20% = 25% of VPIP is a raise.
A sLA with PFR 7% and VP$IP 28%: 7% / 28% = 25% of VPIP is a raise.

I'm not sure if that makes sense.

My idea is, maybe both successful TA and sLa players are playing the # of hands they feel comfortable with, and raising the ones that are in the top 25% or so of those they want to play. So in a sense, both are displaying about the same amount of "relative aggression" per flop seen.

I don't know if this "relative aggression" is a useful category or not. Just thought I would throw it out there and see what people think.

If you are a super-TA with only 12% or 15% VPIP, it doesn't make sense to expect you to PFR 8% or 9% -- that would be most of the hands you play, you raise. You could play that way, of course, but it would be a little strange. (Wouldn't it?)
User avatar
k3nt
Enthusiast (Online)
 
Posts: 6710
Joined: Tue Nov 16, 2004 12:27 pm

Postby gdaviet » Fri Apr 29, 2005 1:05 pm

User avatar
gdaviet
 
Posts: 61
Joined: Thu Apr 14, 2005 7:39 pm

Postby Tiburon » Mon May 02, 2005 10:43 am

"...Every time you cold call, god kills a puppy."
--JJSCOTT2

Read my blog at
User avatar
Tiburon
Semi Pro (B&M & Online)
 
Posts: 1316
Joined: Mon Apr 18, 2005 2:27 pm
Location: South Jersey

Postby Macallan » Tue Dec 20, 2005 10:17 am

Last edited by Macallan on Tue Dec 20, 2005 2:28 pm, edited 1 time in total.
User avatar
Macallan
 
Posts: 13
Joined: Tue Dec 20, 2005 7:20 am
Location: Los Angeles

Postby Tiburon » Tue Dec 20, 2005 11:43 am

"...Every time you cold call, god kills a puppy."
--JJSCOTT2

Read my blog at
User avatar
Tiburon
Semi Pro (B&M & Online)
 
Posts: 1316
Joined: Mon Apr 18, 2005 2:27 pm
Location: South Jersey

Postby Badco » Sun Jan 01, 2006 2:31 am

User avatar
Badco
 
Posts: 17
Joined: Sun Jan 01, 2006 2:11 am

Postby TD » Wed May 24, 2006 9:34 am

This is great work. I would like to a similar analysis throught the pokertracker system. It becomes very important to determine averages in each catagory. As most of my player samples are small, meaningful numbers are rare. Of course, my own sample is large enough but I really don't have much to compare with. I would prefer to adapt a style that is not too far off the median. This way I am somewhat matching the competition and not attracting attention. Therefore, I could adjust my game by piecemeal keeping each stat within optimum range.

For example, this analysis makes me think that my PFA % is too high. I need to selectively call more often.

Keep up the good work.
User avatar
TD
 
Posts: 73
Joined: Sat May 13, 2006 3:59 pm

Postby iStrong » Tue Dec 12, 2006 2:52 am

Sorry to be bumping this thread but i was wondering if this analysis is for 10-max? and if it;s the case, is there a similar analysis for 6-max.

Thx.

P.S.: Great work OP.
User avatar
iStrong
 
Posts: 1
Joined: Tue Dec 12, 2006 2:50 am

Postby Beavis68 » Wed Dec 13, 2006 6:00 pm

User avatar
Beavis68
 
Posts: 3568
Joined: Mon Jul 31, 2006 7:43 pm
Location: double secret probation

Postby DonkiFornication » Thu Dec 14, 2006 5:24 pm

User avatar
DonkiFornication
 
Posts: 731
Joined: Wed Aug 02, 2006 12:53 am
Location: Chi-TownSFobv


Return to PokerTracker and Other Tools

Who is online

Users browsing this forum: No registered users and 0 guests

cron