Which homebrewing package is more accurate?

[frankvw]

After having used Qbrew for many years, I'm now looking at Brewtarget and Beersmith as well. I'm in the process of working out The Ultimate Belgian Dubbel Recipe (yeah right) and decided to give all three of them a spin. The results were wildly different.

So I decided to make a simple test recipe and plug it into Qbrew, Brewtarget and Beersmith 3, using the same parameters as much as possible to eliminate differences in the ingredients databases.

The recipe

• Target batch size: 20L post-boil (not counting losses in the trub, fermenter, etc.)
• Fermentables: 4kg German Pilsener, yield 1.038 / 79%, 2.0°. SRM
• Hops: 28g Cascade 6.5%AA boiled for 60 min.
  
• Yeast: S05

The settings

• Brewhouse efficiency: 70%
• Boiling point: 100°C (i.e. at sea level; no boil altitude / hop utilization adjustment)
• Color estimate: According to Morey
• Bitterness estimate: According to Tinseth
  

The results

• Qbrew: OG=1.044, FG=1.011, IBU=22, Color=3° SRM
• BrewTarget: OG=1.044, FG=1,008, IBU=24.4, Color=3.4° SRM
• Beersmith 3: OG=1.046, FG=1.010, IBU=22.6, Color=3.1° SRM

I realize these software packages have a lot of different features (Beersmit's equipment settings being a good example) but still I can't understand why BS3 comes up with an OG two gravity points higher, while Brewtarget comes up with over two IBUs more bitterness from the same OG.

And this is from the simplest of recipes. With more ingredients the discrepancies only increase further and further. Why is that?

Is there anyone who has used more than one of the above and can give me a pointer on which is the more accurate and why?

 

[kevin58]

In Beersmith everything is dependent on the equipment profile. If you are just plucking one from the provided list or from the add-ons and not using a profile customized to your own system you are just guessing. I have used Beersmith for about 5 years and it is very accurate for me with a customized equipment (and mash) profile.

 

[day_trippr]

"Wildly Different"?

 

[k-os]

"Wildly Different"?

I'd agree that the numbers they're all giving you are not that far off from each other. Is the yeast attenuation set the same for all the softwares (could account for different FG numbers).

 

[ScrewyBrewer]

@frankvw you should post your complete recipe so others can help with the analysis. Personally two or ten IBUs is beyond my taste buds, I wouldn't dwell on that too much.

 

[frankvw]

"Wildly Different"?

For such a simple recipe, yes: 5% difference in OG, 8% difference in IBUs. Add more ingredients and the results vary more with each addition.

 

[frankvw]

In Beersmith everything is dependent on the equipment profile. If you are just plucking one from the provided list or from the add-ons and not using a profile customized to your own system you are just guessing.

Alright. That makes sense. So I created a test profile equipment with the following theoretical (and practically impossible) settings:

• Brewhouse efficiency: 70% (identical to what I used in Qbrew and Brewtarget);
• Mash tun volume, batch volume and boil volume: all 20L;
  
• Mash tun weight, recoverable mash deadspace, mash deadspace losses, top up water for kettle, boil elevation, total whirlpool time, fermenter loss, boil off, cooling shrinkage, loss to trub & chiller and top up water: all zero.

This should essentially replicate Qbrew's and Brewtarget's behavior, i.e. not result in any volume adjustments. Post boil volume as displayed in this equipment profile is now 20 L.

Sure enough, the OG in BS3 is now 1.044 (which matches my own manual calculation and the other two software packages) and the bitterness is 24.2 IBU which matches what Brewtarget comes up with. The world makes sense again!

I have used Beersmith for about 5 years and it is very accurate for me with a customized equipment (and mash) profile.

I guess I'll need to work on that. I've got three 25L plastic buckets with heating elements and a dishwasher pump, and I'm using additional hot liquor to push the beer out of the hosepipes in order to reduce losses, and if I lose too much due to boil-off (I'm at 1800ft and in winter the relative humidity is <20% which promotes evaporation) I simply top up with some boiling water. I suppose Beersmith doesn't work too well for that sort of approach.
Thanks for pointing me in the right direction!!

 

[frankvw]

I'd agree that the numbers they're all giving you are not that far off from each other. Is the yeast attenuation set the same for all the softwares (could account for different FG numbers).

Yes. But given the fact that Brewtarget does not account for differences in wort fermentability due to mash temperature variations, and Qbrew simply puts the FG at 1/4 of the OG no matter what you do, I'm not too worried about that.

 

[pricelessbrewing]

There's a few features missing from brewtarget. Love the project, but it seems to have stagnated the past couple years. I wanted to help, but I don't really know c++ or the tools they're using to be productive...

Fermentability: Beersmith applies a linear equation based on a max fermentability temperature and (I think) the 1st mash step temp within 140-170F. Ideally it should apply different coefficients based on the group of yeast, for example an irish ale yeasts fermentability vs mash temp curve will be different than a diastatic saison that produces glucoamalyse and will bring almost all recipes to ~1.001-1.004.

Have not dug into the other big softwares right now on that topic, brewersfriend/brewfather etc.

Promash/qbrew etc are old, and haven't been updated in years so I would suggest against them.

Small differences in Tinseth formulas, such as using the average boil gravity vs pre or post boil gravity can result in ~10% differences. Classic tinseth uses the average, but imo it would probably be best to use a Riemann sum approach and sum up the incremental IBUs at each minute.

Pretty much every software uses a different approach for whirlpooling / hop steep additions. My favorite atm is the mIBU at aclhemy overlord, https://jphosom.github.io/alchemyoverlord/.

 

[FunkedOut]

I have used BeerSmith but never the others.
I can tell you that in BS, equipment modeling is essential.
The better the model parameter, the better the prediction will be.

In your example, zero loss in the mash tun and kettle (trub and chiller) will force the mash effeciency to be equal to the brewhouse effeciency number you entered.

Zero boil off means no water loss during the boil, no concentration of sugars. Basically just a calculation of sugars for the mash process.

If that is what the other calculators are giving you, then I’d say BeerSmith is a far more advanced modeling package with many more parameters seen in real world processes.

You’ve seen how easy it is to tune them out.
They’re there for you when you want to include them and better approximate your process.

Nothing has helped me completely wrap my head around the entire brewing process in such detail as much as building my BeerSmith equipment profile.
The first go around takes some time.
But as you start to refine each parameter with brew day measurements, your predictions get closer and closer.

My equipment profile is now as tight as I can measure.
That was a process that in and of itself was and enjoyable process for me.

 

[beernutz]

For such a simple recipe, yes: 5% difference in OG, 8% difference in IBUs. Add more ingredients and the results vary more with each addition.

Wildly different would be the case if you made 3 beers which exhibited the exact estimates of OG, FG, IBUs, and SRM that each software package predicted and you could tell them apart. My guess is you couldn't. I know I couldn't.

I still use BS2 and have no experience with the other two but I know I am able to adjust my equipment profile to take into account differences between the estimates BS provides and my actual results, at least for OG and FG. If you can tell the difference between 3.0 and 3.4 SRM or between 22 and 24.4 IBUs you may be a supertaster.

I am reiterating what's already in this thread but tinkering with my equipment profile to make BS's estimates more closely match my results is probably the most satisfying part of using that software.

 

[kh54s10]

Those "wildly" different numbers are calculations. The differences, IMO, are so insignificant that side by side each brewed as identically as possible and a blind taste test no one could possibly taste or see any difference.

 

[frankvw]

There's a few features missing from brewtarget. Love the project, but it seems to have stagnated the past couple years. I wanted to help, but I don't really know c++ or the tools they're using to be productive...

I wanted to help, too, if only by adding the BJCP2015 to the database, but I'm not even getting a response. So yes, BT seems to be somewhat dead in the water.

I agree with your other comments. Tnx!