Colorcoding Stars annoyingly SLOW

[08-15am]

I try to use the colorcoding on Stars but although I have a very propper machine, which processes roughly 120-150 hands/sec when notetaking, the colorcoding is extremely slow. It takes almost 20 hours to do the colorcoding although I have very simple colordefinition:
1) Aggro (WWSF > 51 Sample size 1000) using the HM2 stats
2) Fish (VPIP > 40 sample size 30) using the HM2 stats
3) Loser (Winrate < -3 sample size 5000) using the HM2 stats
4) Breakeven ( -3 < Winrate < 3 sample size 5000) using the HM2 stats
5) Winner ( Winrate > 3 sample size 5000) using the HM2 stats

I have roughly 20M hands in my DB but this is pretty useless if the job blocks my machine for one complete day.

Can you please advice how to speed this up!

[HossendeHaai]

A 20kk DB will lead to long processing times on any PC with any settings. I don´t see how you choose to use so many datamined hands, accept the notewriting to take multiple days, but find color coding within a day to be useless.

You could try the things described here: http://wiki.assaultware.com/NoteCadd...provement.ashx. Also the warning not to use HM2 stats for color coding seemed pretty clear to me.

NC.png

I´d try using the Compiled Statistics instead.

[08-15am]

A 20kk DB will lead to long processing times on any PC with any settings. I don´t see how you choose to use so many datamined hands, accept the notewriting to take multiple days, but find color coding within a day to be useless.

You could try the things described here: NoteCaddy speed improvement - Assaultware Wiki. Also the warning not to use HM2 stats for color coding seemed pretty clear to me.

NC.png

I´d try using the Compiled Statistics instead.

Thanks for the subtle hint. I use HM2 stats because of this advice:

It seems the sample size isn't being grabbed from the database correctly. You can either:
-remove the sample size filter for WWSF
-wait till the next release where this will be fixed (today or tomorrow)

Thank you for letting me know and apologies for the inconvenience

AND

I don't have any control over the stats hm2 returns. Compiled stats aren't always up to date. Also, NC grabs them per game type/size and I don't think the opponents tab splits it like that hence a smaller sample size would make sense. You can try the "win rate" stat in misc under hm2 stats.

so will Compield stats work and be faster?

[HossendeHaai]

As announced in the message you quoted an update that is supposed to fix the issue has been released: Minor Release notes (NoteCaddy change log). So I guess it's fixed.

Compiled stats will work, and I never had any speed issues with it. The only thing is that you have little control over the hands the stat is based on, but that seems inevitable. To have full control, you'd have to replicate the stats you want as NC definitions.

[08-15am]

The only thing is that you have little control over the hands the stat is based on, but that seems inevitable.

What does that mean specifically. Could you please explain this to me?

[HossendeHaai]

I haven't experienced it myself, but I think this explains it pretty well: Understanding why NoteCaddy stat values may not match HM2 stat values - Assaultware Wiki.

[08-15am]

I haven't experienced it myself, but I think this explains it pretty well: Understanding why NoteCaddy stat values may not match HM2 stat values - Assaultware Wiki.

When I swithced the color definitions back to compiled stats, this worked fine when testing for Fish / Aggro but
the winrate dependant color definition failed. The test replies with a winrate of 0 for all players.

Is this just a testing issue or is the bb/100 in compiled statistics not working?
Could someone please tell me?

Thanks

[sreticentv]

When I swithced the color definitions back to compiled stats, this worked fine when testing for Fish / Aggro but
the winrate dependant color definition failed. The test replies with a winrate of 0 for all players.

Is this just a testing issue or is the bb/100 in compiled statistics not working?
Could someone please tell me?

Thanks

I just tried winrate and it works fine in test, seek, etc

Found stats for : Holdem game size Medium Hands: 3 VPIP: 23.13 PFR: 17.73 Agg%: 36.70 AF: 2.89 BB/100: -0.02
OUTPUT: FAIL - reason: Winrate too small
Found stats for : Holdem game size Full Hands: 19 VPIP: 21.05 PFR: 15.79 Agg%: 50.00 AF: 8.00 BB/100: 281.58
OUTPUT: FAIL - reason: Winrate too big
Found stats for : Holdem game size HeadsUp Hands: 821 VPIP: 49.09 PFR: 40.19 Agg%: 38.27 AF: 2.70 BB/100: -7.24
OUTPUT: FAIL - reason: Winrate too small
Found stats for : Holdem game size Medium Hands: 117 VPIP: 25.64 PFR: 23.08 Agg%: 7.14 AF: 4.00 BB/100: 4.99
OUTPUT: SUCCESS
Found stats for : Holdem game size HeadsUp Hands: 37 VPIP: 37.84 PFR: 32.43 Agg%: 12.50 AF: 6.00 BB/100: 67.41

[08-15am]

I just tried winrate and it works fine in test, seek, etc

Found stats for : Holdem game size Medium Hands: 3 VPIP: 23.13 PFR: 17.73 Agg%: 36.70 AF: 2.89 BB/100: -0.02
OUTPUT: FAIL - reason: Winrate too small
Found stats for : Holdem game size Full Hands: 19 VPIP: 21.05 PFR: 15.79 Agg%: 50.00 AF: 8.00 BB/100: 281.58
OUTPUT: FAIL - reason: Winrate too big
Found stats for : Holdem game size HeadsUp Hands: 821 VPIP: 49.09 PFR: 40.19 Agg%: 38.27 AF: 2.70 BB/100: -7.24
OUTPUT: FAIL - reason: Winrate too small
Found stats for : Holdem game size Medium Hands: 117 VPIP: 25.64 PFR: 23.08 Agg%: 7.14 AF: 4.00 BB/100: 4.99
OUTPUT: SUCCESS
Found stats for : Holdem game size HeadsUp Hands: 37 VPIP: 37.84 PFR: 32.43 Agg%: 12.50 AF: 6.00 BB/100: 67.41

You mean you tried winrate or you tried compiled statistics bb/100 ?
Winrate works for me too but bb/100 in compiled doesnt result any bb/100 .. it seems to be zero for all players.
According to the warning to not use HM2 statistics and in the hope to speed up color coding I thought it is recommended to switch back to compiled stats which wasnt working in the first place.

Could you please explain to me if it is quicker to use compiled stats instead of HM2 stats?
And what stat I can use to label winning / losing player.

Thanks in advance!

[sreticentv]

It is much quicker to use compiled stats. If it's showing 0 for all players then there is something wrong with the data. You can try to make a new database which will likely fix it