Rounding Counts

[Krainn]

Hey guys,

So I've recently started playing with CV and I love it! However, I've always played on the safe side with counts, meaning that if I've discerned the TC as being 1.5 or 1.75, I would play as if the count was not yet 2 and was still 1. I noticed that CV rounds it up, so mathematically I'm guessing that its better to play as if the TC is 2 if I calculate a 1.5 instead of playing as if the TC is 1? Any clarification or qualification is appreciation.

Thanks!

 

[sagefr0g]

if i remember correctly, one can set the software to round up or down. perhaps so it could fit ones certainty equivalent sort of thing. just guessing on my part.
far as rounding up being better, here's a guess, it may or may not be better, but rounding up would put one in a position wherein realization of opportunity risk would be lower.

 

[DSchles]

Hey guys,

So I've recently started playing with CV and I love it! However, I've always played on the safe side with counts, meaning that if I've discerned the TC as being 1.5 or 1.75, I would play as if the count was not yet 2 and was still 1. I noticed that CV rounds it up, so mathematically I'm guessing that its better to play as if the TC is 2 if I calculate a 1.5 instead of playing as if the TC is 1? Any clarification or qualification is appreciation.

Thanks!

No, that isn't right. CV does NOT automatically just round up; rather, if you choose flooring of true counts (which you should), all calculations will round down to the next integer. So, TC of 1.75 is TC = 1. TC of -2.2 is TC = -3.

Don

 

[Taff]

Krainn play it as you were. From what I recall you're saving hard for your initial bankroll. Rounding down is the correct move. Variance would be lower as would risk to B.R.

 

[gronbog]

Rounding down is the correct move.

Just to be precise here, "Rounding Down" could be interpreted in two ways

• Chop off the fraction -- this is more precisely called truncating -- -1.x becomes -1, 1.x becomes 1
• Adjust to the next lower integer -- this is more precisely called flooring -- -1.x becomes -2, 1.x becomes 1.

As Don says, flooring is what you should be doing. I think that's what Taff was trying to convey, given his comments re: variance and risk.

 

[Taff]

As Don says, flooring is what you should be doing. I think that's what Taff was trying to convey, given his comments re: variance and risk.

Sorry yep. Should have been clearer. Adjust down to next lowest integer or 'floor it'.

 

[Krainn]

Thank you Taff, gronbog, and Don. I greatly appreciate the helpful nature of this website's community! Its been of tremendous aid to me.

 

[xengrifter]

As Don says, flooring is what you should be doing.

Why "should"?
I thought it's long established that there is no appreciable difference in results obtained from flooring, rounding or truncating?
Rounding is easiest for many.

 

[gronbog]

I don't want to answer categorically without specifics to back me up (which I don't have handy), but will say that I believe that flooring is the most risk averse of the 3 and that truncation results in a double-wide betting bin around TC=0 which dilutes the data for the most frequent true counts: -1 < TC < 1

 

[sagefr0g]

it's interesting that in the link below, it's stated that, "Truncation is a bit inferior due to this large peak at TC zero. When you have so many hands all identified as a count of zero, the count is less precise and an index of zero less valuable. However, the overall effect on advantage is not high."
https://www.blackjackincolor.com/truecount4.htm
point being imho, ones count may be more precise depending on how one determines the integer, but not so much the advantage obtainable.
edit: seems to me, if those bins about tc=0 >-1 & <+1 are ignored,(eg. especially if tc >0<1 not treated as tc=+1) that opportunity risk might become meaningful, since less tc= +1 bets would be made when perhaps they should be. end edit

 

[xengrifter]

it's interesting that in the link below, it's stated that, "Truncation is a bit inferior due to this large peak at TC zero. When you have so many hands all identified as a count of zero, the count is less precise and an index of zero less valuable. However, the overall effect on advantage is not high."

Yes, "not high" as in microscopically insignificant.

 

[KewlJ]

"Microscopically insignificant" like a lot of other issues involved with counting?

 

[xengrifter]

"Microscopically insignificant" like a lot of other issues involved with counting?

Yes but unlike other issues, the long-running debate over truncate, floor or round is meaningless ...
... except that rounding is easier/faster - the one meaningful determinant.

 

[KewlJ]

Yes but unlike other issues, the long-running debate over truncate, floor or round is meaningless ....

I would argue that the long running debates over some "other issues" are just s meaningless. And if not meaningless....then your "microscopically insignificant" certainly seems appropriate.

 

[21forme]

I would argue that the long running debates over some "other issues" are just s meaningless.

Absolutely agree. Yet that meaningless debate continues to consume an entire forum year after year.

 

[KewlJ]

Absolutely agree. Yet that meaningless debate continues to consume an entire forum year after year.

In keeping with my agreement not to discuss that particular site, I will only quibble with your use of the word 'forum'.

According to Oxford Dictionary, 'forum' is defined as "a place, meeting, or medium where ideas and views on a particular issue can be exchanged". When the person running the "meeting" or in control of the "medium", only allows a singular view point and banishes or silences those with an opposing view, then you don't so much have a "forum" where ideas can be shared and discussed, as a private discussion among a small, specifically selected, group of members, all with the same basic view point.

 

[Fenix]

There are three options, rounding, truncating, and flooring. If I choose rounding, does it ever round up? i.e. does a TC of 3.8 become 4? Also, other than flooring being more risk averse, is there any reason against rounding? It seems easier to me, but before I get to memorizing a bunch of things I want to make sure I'm not overlooking something important.

 

[xengrifter]

a small, specifically selected, group of members, all with the same basic view point.

Sounds like a cult.

 

[gronbog]

Also, other than flooring being more risk averse, is there any reason against rounding? It seems easier to me

For rounding you need to know whether the fractional part of the true count is < 0.5 or >= 0.5 in order to make the proper adjustment.

For flooring, you only need to know that there is a fraction. You don't need to know what it is. This lends itself to being able to think about the true count as a series of running count thresholds, eliminating the need for any division at all at the table.

For me, this makes flooring easier than rounding.

 

[Fenix]

Thank you Gronbog .