32 replies to this topic

[chuck_d]

I don't know that I'd call (OG-FG)/OG a tool and nobody's trying to deny it to anybody. It's just that it's not very useful. I suppose but I think the real power of brewing software is making calculations that I don't know how to make. Such as what temperature my strike water should be. But it's a moot point anyway.

It's not (OG-FG)/OG, that's how you'd calculate attenuation, not predict FG. It's something more along the lines of:OG-((Attenuation High + Attenuation Low)*OG/2), but I just wrote that off the top of my head so nobody use that, if you want the formula I can dig it out of my spreadsheet.You can look up how to make strike water calculations, they are right in Palmer's HTB on p. 170, and then use your calculator & pencil :unsure:Or: https://dieseldrafts...t/Formulas.htmlInitial Infusion EquationStrike Water Temperature TW = (0.2/R)(TS-TG)+TS where TW = temp of strike water, R = mash thickness in quarts per pound or liters per kg, TS = desired mash step temp, TG = temperature of the grains, temps can be C or F but use 0.41 for Cthis does not account for heat lost to tunPersonally, I think having all these numbers just presented right in front of you while composing a recipe is useful. Different strokes...

[MtnBrewer]

It's not (OG-FG)/OG, that's how you'd calculate attenuation, not predict FG.

It's the same thing. That's my point...that's all brewing software does to "predict" FG. There's no magical equation in there that tells you what your FG should be. They all just use (OG - FG)/OG. So the best approach is to decide what you want your FG to be and design your recipe and process accordingly.

[chuck_d]

It's the same thing. That's my point...that's all brewing software does to "predict" FG. There's no magical equation in there that tells you what your FG should be. So the best approach is to decide what you want your FG to be and design your recipe and process accordingly.

Exactly, it predicts your FG based on (in my formula) the average attenuation of the yeast strain you choose, thus giving you a tool to design a recipe once you have decided what you want your FG to be. I don't know what you mean by magical equation exactly, but it seems like we're finally in some sort of agreement

[MtnBrewer]

Exactly, it predicts your FG based on (in my formula) the average attenuation of the yeast strain you choose, thus giving you a tool to design a recipe once you have decided what you want your FG to be. I don't know what you mean by magical equation exactly, but it seems like we're finally in some sort of agreement

As long as you understand the limitations of the "tool" that's fine. But many people don't and they come here going, "$BREWING_SOFTWARE predicted that my FG should be 1.012 and it's 1.018. Why?"

[Sidney Porter]

could FG be perdicted with a formula?We know the posted range for the yeast.We know the mash temp.We know the grain bill, adjuects, sugar, etc.We know the fermentation temp.In theory we could know the number of viable cells pitched.In theory we could know the amt of disolved o2.It is my understanding that each manufacture use the same grain bill to come up their numbers. So they would probably need to be an adjustment between white labs and wyeast.

[chuck_d]

As long as you understand the limitations of the "tool" that's fine. But many people don't and they come here going, "$BREWING_SOFTWARE predicted that my FG should be 1.012 and it's 1.018. Why?"

Are we back on this again? If someone comes in here and asks that, it's a perfect opportunity for experienced and knowledgabled people like us to help them understand. This argument could be made for any of the calculations in brewing software. "Why was my color different than predicted?" "How come I have a different OG?"

[chuck_d]

could FG be perdicted with a formula?We know the posted range for the yeast.We know the mash temp.We know the grain bill, adjuects, sugar, etc.We know the fermentation temp.In theory we could know the number of viable cells pitched.In theory we could know the amt of disolved o2.It is my understanding that each manufacture use the same grain bill to come up their numbers. So they would probably need to be an adjustment between white labs and wyeast.

Yes, a prediction is just an estimation. See my formula above. Attenuation shouldn't be affected by by cell count or O2 as long as both are reasonable. The exact same thing could be said of color, which apparently nobody seems to have a problem with, the prediction of that is just an estimation. The big difference there is that two beers of the same SRM value can actually be different colors! Two beers with the same attenuation & OG will have the same FG.

[MtnBrewer]

Are we back on this again? If someone comes in here and asks that, it's a perfect opportunity for experienced and knowledgabled people like us to help them understand. This argument could be made for any of the calculations in brewing software. "Why was my color different than predicted?" "How come I have a different OG?"

Wasn't aware that we'd moved on...sorry.The difference is that the other calculations are correct and so the answer usually is that they aren't using the right mash tun thermal mass or something like that.

[chuck_d]

Wasn't aware that we'd moved on...sorry.The difference is that the other calculations are correct and so the answer usually is that they aren't using the right mash tun thermal mass or something like that.

What do you mean they are correct? Color is just BS, I don't even think any pro-brewers bother to predict color (at least that's what Stephen Holle's book leads me to believe) and as I mentioned you can have to beers with the same measured SRM that are different colors (not predicted SRM, measured SRM). For IBU there are 3 different systems that provide vastly different numbers. Even predicting OG is an approximation, since we often don't know what the dry yeild of our grains are, not to mention that those numbers are determined by a congress mash. AA% are averages for the crop, they deterioate in storage and aren't updated by the homebrew shops at the time of sale (as they should be if we want our IBU calculations to be correct, otherwise if I bought the same hops a month later they'd have a lower AA%). ALL of it just an estimation to give you an idea what you're going to be getting. So the three materials we using we have estimated attributes (our grains, hops & yeast), the difference being that FG calculations are by definition, whereas color can definitely be wrong, and bitterness we can't even agree on what the most accurate formula (see the current Rager vs Tinseth discussion).

Edited by chuck_d, 10 April 2009 - 10:21 AM.

[MtnBrewer]

What do you mean they are correct? Color is just BS, I don't even think any pro-brewers bother to predict color (at least that's what Stephen Holle's book leads me to believe) and as I mentioned you can have to beers with the same measured SRM that are different colors (not predicted SRM, measured SRM).

SRM can be calculated pretty accurately. If there's a flaw in the SRM system itself, it's not the fault of the computation that arrives at that number.

For IBU there are 3 different systems that provide vastly different numbers.

One of which is probably pretty accurate for your system. At least it attempts to model utilization.

Even predicting OG is an approximation, since we often don't know what the dry yeild of our grains are, not to mention that those numbers are determined by a congress mash.

OG can be calculated precisely given good data. Again, garbage in is not the fault of the calculation but the person entering the data.

AA% are averages for the crop, they deterioate in storage and aren't updated by the homebrew shops at the time of sale (as they should be if we want our IBU calculations to be correct, otherwise if I bought the same hops a month later they'd have a lower AA%).

Which can be modeled fairly well.

ALL of it just an estimation to give you an idea what you're going to be getting.

Right. That's the difference between them and (OG-FG)/OG.In any case, FG doesn't really need to be calculated because I see it as an input to the process not an output. Once you've brewed the beer, you can simply measure FG, so there's no need to calculate it.

[stellarbrew]

I think an FG prediction is inherently less accuarate than most of the other calculated predictions we make. It goes soley by the OG and the published average attenuation of the yeast. Does one interpret that published average attenuation to mean that on average, this yeast has been demonstrated to attentuate this percentage of attenuable sugars? Or is it all sugars, attenuable and unattenuable? After all both attenuable sugars and unattenuable sugars contribute to the OG. The amount of unattenuable sugars of course varies considerably, depending on mash rest temperature(s), grain bill specifics, mash pH, etc. Is that published average attenuation an average over all possible fermentation temperatures, levels of pH, levels of oxygenation, pitching rates, and wort original gravities? Or does it represent an average rate of attenuation arrived at by the yeast manufacturer within a narrow range of testing conditions? I just think that the whole process of determining how all of those factors, in combination or separately, will affect the actual attenuation a yeast will give you for a particular beer is much too complicated to derive any remotely reliable empirical formula for prediction. There are just too many variables, and the variables are just too complex.I find an FG calculation somewhat useful in formulating recipes, in that it gives me some idea, at least relatively speaking. But I know from my own experience that I can nail my target OG, yet still get an actual FG a which is pretty far away from what was predicted by the average attenuation of the particular yeast I used. What I do is make mental note of the discrepency and try to keep in mind for predicting results with future recipes. But mainly I just keep in mind that the FG predicted is not usually a very reliable number.

[chuck_d]

That's my point though, it's not. Take for example the 41-45 minute addition of hops. Rager will predict 50% more utilization than will Garetz (based on a quick look at the tables in the link I posted in post #7). Taking WLP001 73-80% averaged is 76.5%, but let's assume that we do go outside the range and achieve 82% AA. The difference in final gravity b/w the predicted and measured values will be 30%, a smaller difference than the Rager/Garetz difference. And that's just choosing which formula you want to use. If we assume 85% AA then the difference will be 56%, on par with the difference in formulas for bitterness. The reason we don't notice the fact that predictions for color, IBU, etc. are off is because we don't measure them.

[stellarbrew]

That's my point though, it's not. Take for example the 41-45 minute addition of hops. Rager will predict 50% more utilization than will Garetz (based on a quick look at the tables in the link I posted in post #7). Taking WLP001 73-80% averaged is 76.5%, but let's assume that we do go outside the range and achieve 82% AA. The difference in final gravity b/w the predicted and measured values will be 30%, a smaller difference than the Rager/Garetz difference. And that's just choosing which formula you want to use. If we assume 85% AA then the difference will be 56%, on par with the difference in formulas for bitterness. The reason we don't notice the fact that predictions for color, IBU, etc. are off is because we don't measure them.

That's a fair point. I had never really looked at any of the other hop utilization models before. After reading the thread you mention, and looking through descriptions of those various models you posted, I am beginning to conclude that both Rager and Garetz are more primitive, less accurate models than Tinseth. I had never actually put too much stock in the accuracy of any of the hop utilization models, and I figured that I was at best getting a relative idea of bitterness based on past calculations. I perceived that the utilization calculation accuracy might be on a par with attenuation/FG calculation accuracy. After reading more now, I am of the opinion that if you use the Tinseth model you will probably get a higher order of accuracy for predicting hop utilization than the order of accuracy you can expect for predicting attenuation. At least one example to validate the accuarcy of Tinseth is Denny's example. He has actually had the tests done and has found in those case that Tinseth is within 1 IBU of being accurate.