Baker's yeast quantity prediction model - please compare to your results

[TXCraig1]

EDIT: The latest yeast quantity prediction model can be found at Reply 261 at https://www.pizzamaking.com/forum/index.php?topic=26831.msg393271#msg393271 (it replaces Reply 188 at http://www.pizzamaking.com/forum/index.php/topic,26831.msg349349.html#msg349349); for a version of the chart with the yeast percents stated out to three decimal places, see Reply 202 at http://www.pizzamaking.com/forum/index.php?topic=26831.msg355933#msg355933; for a further expanded version, see Reply 399 at https://www.pizzamaking.com/forum/index.php?topic=26831.msg511590#msg511590; for a 48-hour chart, see Reply 406 at https://www.pizzamaking.com/forum/index.php?topic=26831.msg511818#msg511818; beyond 48 hours, see Reply 202 at https://www.pizzamaking.com/forum/index.php?topic=26831.msg355933#msg355933

The model works the same way as my sourdough starter predictive model (http://www.pizzamaking.com/forum/index.php/topic,22649.0.html) except it's for baker's yeast (ADY, IDY, and CY). It's still in the developmental stages, so please don't rely on it yet.

What I'm hoping you will do at this point is tell me where the predictions differ from your proven results. For example, it predicts 0.05% (0.048% column on the chart) IDY fermented at 70F will take 12 hours. Maybe you've done something similar, and it only took 8 hours - please post what you did - your yeast%, temperature(s), and time(s), so I can fine tune it.

Thanks in advance.

[deb415611]

These charts are awesome Craig, thank you. 

Yeast %'s are what I struggle with most.    I'll be using this and will let you know variances if any.

[JimmyG]

Great idea Craig. Is there a regression or polynomial equation that you are currently working from? If so, would you be willing to share your current equation? Just curious is all.
Jim

[Pete-zza]

Craig,

To add or expand upon JimmyG's question, are you using a reference standard of some sort, such as the doubling of the dough--bulk or individual dough balls?

Peter

[TXCraig1]

I started here: http://www.pizzamaking.com/forum/index.php/topic,26602.0.html.  Converted it to yeast % based on 60%HR and decided to use 25C as my baseline model. Why 25C? 25C = 77F which is the temperature I keep my house and I like the way it sounds. 7 is a lucky number, right? So two 7's must be twice as lucky - all of which of course means it was a SWAG (Scientific Wild Ass Guess). Also, when I first plugged it into the model, some points I felt good about around the table were reasonably close. Basically, the model starts with this curve and then uses a growth model to extrapolate across the temperature range.

I had a very hard time finding any good data on growth rate at temperature for strains of  S. cerevisiae used in baking, so I used the model proposed by Ganzle et al. (1998) - http://aem.asm.org/content/64/7/2616.full.pdf. wiggle fit for baker's yeast as best I could with the data I have. I started with Salvadó et al. (http://aem.highwire.org/content/77/7/2292.full.pdf) and Serra et al. (http://oatao.univ-toulouse.fr/1556/1/Serra_1556.pdf) and then hand adjusted the curves and let Excel Solver fit them to the Ganzle model until I hit or was reasonably close to data I feel good about at several points around the chart you see above. I checked it against some of the data Peter pulled for me a while back and it was generally in the ball park of the points I checked.

If I get some feedback, I can easily adjust both the baseline and the growth model.

Ganzle model:
growth rate at temperature T = a ∙ x^b ∙ e^-c∙x

x = T_max - T

T_max is the maximum temperature at which the yeast will grow.

In the model right now,

Tmax = 45C
a= 0.02645608
b= 2.037020784
c= -0.198964236

[TXCraig1]

The fermentation time in the charts is total bulk + balls. Reballing, punching down, and other optional activities that would serve to artificially extend fermentation will not work with this model - or at least you would have to allow for extra time from what is shown.

[JimmyG]

Craig I'm guessing the parameter e^-c∙x stands for the natural log raised to a time constant or a doubling time or sorts, is that correct? The authors didn't define their terms very well.  In any case very cool and a great idea you are developing.

Jim

[TXCraig1]

Jim, e^x is the exponential function (natural log would be the inverse - log base e or ln). I don't know that it represents anything particular - rather the whole function including that term appears to be a good generalized model for microbial growth.

CL

[JimmyG]

ex is the exponential function (natural log would be the inverse - log base e or ln)

Yeah thats right, I was looking in my statistics text book at the logistic regression function (1/1+e^-x), and said natural log for some reason, but yeah, e^x is absolutely the exponential function. But yeah I would guess that term probably stands for some sort of exponential doubling time or growth of the bacteria. 
Jim

[TXCraig1]

Here is my growth model along with Serra and Salvado for comparison. Remember that while the Serra and Salvado charts are both for particular strains of S. cerevisiae, neither is a strain found in baker's yeast. My model was built from these two and wiggle fit to various known baker's yeast data points. The more data people can give me, the better I can fit the curve and the more accurate the chart will be.

[JHutchins]

Craig, what do you consider to be the fermentation time? The time to double in volume?

[TXCraig1]

Craig, what do you consider to be the fermentation time? The time to double in volume?

Ready to bake.

[Barry]

Hi Craig,

This is a fantastic tool!!  Thank you for sharing.

Best wishes

Barry in Cape Town

[JHutchins]

Ready to bake.

Isn't that very subjective? At what point do you consider it ready to bake?

I use the seed method to measure dough rise (http://www.pizzamaking.com/forum/index.php?topic=6914.0). Pizzas seem to turn out fine between 1 1/4" and 1 1/2" which is a doubling and tripling in volume. The fermentation time for each is very different but they are both ready to bake.

[TXCraig1]

Isn't that very subjective? At what point do you consider it ready to bake?

I use the seed method to measure dough rise (http://www.pizzamaking.com/forum/index.php?topic=6914.0). Pizzas seem to turn out fine between 1 1/4" and 1 1/2" which is a doubling and tripling in volume. The fermentation time for each is very different but they are both ready to bake.

I see your point. I was trying to avoid the opposite condition - where I specify 2X and users say "what if I don't bake at 2X?" If I had enough data, I think I would prefer this method as it should guide the data to the mean as opposed to me setting the mean at an arbitrary figure - as you noted you bake between 2X and 3X.

Specifying 2X is probably the more objective way to do it - particularly given the limited feedback.

To be honest, I'd be happy if people woul give me some feedback either way. 

[parallei]

I've never paid much attention to an increase in volume with dough balls, though maybe I should.  I do pay attention with proofing with bread.  I look at the bottom of the dough and its overall condition.  I suspect many just do this.

Perhaps folks could attach a photo of their dough that shows when they've deemed it "done" and ready for the bake.

For fermentations done in the fridge, it might also be helpful to see and out of the fridge and after "x" hours at room temp.  Mine change a bit over the two hours or so at room temp.

[scott123]

Craig, yeast activity is impacted by variables other than yeast%, temperature and time. These other variables include:

hydration
salt
oil/fat
sugar
yeast age
dough ball size
gluten development
altitude
water chemistry (level of chlorination especially)
container material and thickness (conductivity if ambient and dough temps vary, along with heat dissipation from fermentation)
flour chemistry (enzyme activity/damaged starch etc.)

Some of these variables might have lesser impact than others, but, collectively, they put the usefulness of member feedback in question.  If, say, you wanted to base the model on IDY NP, then the use of the same flour, typically tight hydration and salt levels, along with very similar dough ball sizes might help you fine tune the model for NP bakers, but asking everyone to give you yeast%, temp and time numbers isn't going to give you useful data, imo. I could easily make two doughs with identical yeast %, ferment them at the same temp, but, by modifying the other variables, double or half the time when they're 2x-3x volume (aka 'ready').

[TXCraig1]

Craig, yeast activity is impacted by variables other than yeast%, temperature and time. These other variables include...

Some of these variables might have lesser impact than others, but, collectively, they put the usefulness of member feedback in question.

I'm well aware of that; notwithstanding, I disagree with your conclusion. In my experience, if those factors are held within relatively "normal" ranges they* will not have a particularly meaningful impact on fermentation time – not even collectively – certainly nothing outside the scope of this project. Actually, I'd say you are completely wrong about the usefulness member feedback, the more data I get, the more the noise from the factors you noted is cancelled out. And give me a little credit; I know an outlier when I see it.

If, say, you wanted to base the model on IDY NP, then the use of the same flour, typically tight hydration and salt levels, along with very similar dough ball sizes might help you fine tune the model for NP bakers, but asking everyone to give you yeast%, temp and time numbers isn't going to give you useful data, imo.

IMO, you're wrong. With a little member feedback, this could be plenty accurate to peg a starting point to work from for all sorts of dough styles – NP, NY, American, etc. That's the goal – to help people find a starting point – not to tell them exactly how to formulate their dough.

I could easily make two doughs with identical yeast %, ferment them at the same temp, but, by modifying the other variables, double or half the time when they're 2x-3x volume (aka 'ready').

Double or half the time? Come on now, who are you trying to kid? Perhaps with some fringe, outlier formulations that nobody would ever use, but even then I doubt it – not even by doing something totally goofy with the pH or sugar. When it comes to fermentation within typical parameters, nothing even comes close to the impact of yeast quantity and time. If I'm wrong, let's see the formulas and pictures to prove it.



*Yeast age? Seriously? How many active members do we have that don't pay attention to the condition of their yeast? Any member who does not think it is important to use yeast that is in date – please don't reply to this thread.    And unless you have some bizarrely chlorinated water that the EPA should probably know about, it's not going to hurt your yeast.

[scott123]

Craig, define "normal." 45% hydration is perfectly normal for a cracker crust, while 75% and higher is normal for some styles of Sicilian.  Hydration dictates water activity and water activity is a major component of yeast activity.

Malted flour is 'normal' for NY. Unmalted flour is 'normal' for NP.  Because of the lack of enzymes/enzyme generated sugar, you're looking at two different rates of fermentation.

0% sugar is normal for NP, but I've seen American frequently go as high as 4% sugar.  Huge difference.

Active members who work with yeast tend to overwhelmingly work with bottles.  Yeast packets have dates that tend to creep up on you. Bottles are good for a year or two.  I would be shocked if any of the active members working with bottled yeast knew what the expiration date on their yeast was. I don't.   At the same time, though, it's very common for members to go through these bottles of yeast slowly.  I've had a bottle of yeast for a year.  Refrigerated bottled yeast doesn't die after a year, but it loses some of it's punch. 

Based upon my own observations, I'm being very conservative in saying that yeast loses at least 15% of it's vitality over the course of a year.

Do you really feel that the EPA has rigid guidelines on how much chlorine is allowed in tap water?  Really, Craig, the government?  They have relatively tight guidelines on bacteria counts, but, whatever methods are used to achieve those bacteria counts have a lot of leeway. I'm sure you're aware of the multiple methods in which water is purified. You can have communities using other methods than chlorine (ozone, uv light, etc.), resulting in water with no chlorine, and, for those communities that do use chlorine, it has an incredibly large range.  I don't need water reports to be aware of this.  I can taste it in the places I visit.  Some places have lots of chlorine, some less, some none.

Yeast is a fungus and chlorine is a fungicide. Varying amounts of chlorine (from zero % to water that tastes and smells like a swimming pool) while promote/inhibit yeast growth.

Speaking of fungicides, salt is a fungicide.  1.75% salt is 'normal' for NY and 3% salt is 'normal' for NP.  Do you really feel that a difference in 1.25% isn't going impact yeast growth perceptibly?

I've seen what hydration can do to yeast activity.  I don't need studies to prove that.  I've not worked with unmalted flours, but I have been tracking typical unmalted yeast quantities, and they are, across the board, higher than quantities required for malted flours. As I've said, I've seen yeast lose it's umph over time. I've also worked with chlorinated water and unchlorinated water and I've witnessed the effects of chlorine. You want proof? My memories are proof.

Do any of these variables impact dough as much as time or temp?  Individually, no, but, together, it all adds up.

You give me a 48 room temp NP dough, and, with a 'normal' formula, I can at the same temp, give you a dough that's ready in 24 or 96. I might be able to do it with just hydration (45% and 75%), but that could be a bit too ambitious.  Hydration + chlorinated water/non chlorinated + 0%/4% sugar + unmalted/malted flour + 1.75%/3% salt + fresh/1 year old yeast- well, that's a piece of cake.

[TXCraig1]

Craig, define "normal." 45% hydration is perfectly normal for a cracker crust, while 75% and higher is normal for some styles of Sicilian.  Hydration dictates water activity and water activity is a major component of yeast activity.

I put quotes around it precisely because I don't think defining it is necessary. It's just a distraction that will generate off-topic questions and discussion like this. If you insist, it would be something like this: 65% +/- 10%HR, 2% +/- 2% sugar, 2% +/- 2% oil, 2% +/- 1% salt, tap or bottled water, malted or un-malted flour, and yeast (ADY or IDY) 1 year old or less or fresh CY. That should be no surprise to you. I bet this represents 99% of what is posted in the NY and NP threads. That's why I didn't define it. You know I'm not talking about cracker crusts. I'm talking about normal flat types of pizza – the stuff most people here want to make.

I think maybe you've missed the whole point of this exercise. No less than 25 people have asked me to come up with a chart similar to the SD predictive chart but for baker's yeast, and just like that chart, it's intended to be a starting point – not a guaranteed solution to every formulation question. Today if someone has a particular dough and wants to increase or decrease the fermentation time or maybe try something new from scratch, they could look through 100's of posts and never find the yeast information they need to get started. This chart is intended to simplify that process – to give people a place to start. They will still need to experiment and fine tune things.

I've worked with malted and un-malted flour; tap water, RO water, and bottled water; salt from 1% - 3.1%; sugar from 0% - 4%; oil from 0% - 3%; ADY and IDY – some not perfectly fresh; and I believe that while these ranges could introduce some noise, given that I get even a dozen or so responses, I can make a useful chart. But I need some data. When you look through the pages here, it seems that almost everyone cold ferments. The variability afforded in cold fermenting (you can let the same dough go for a wide range of days as Tom proves over and over) makes that data less valuable but not worthless. None of this withstanding, I think if the model is well tuned above refrigerator temps, it will also provide a reasonable baseline for cold fermenting.

Statistics are a beautiful thing.  They have a wonderful way of teasing out the data you want from a jumbled mess. All these things you bring up are just noise and noise cancels itself out with enough data allowing the signal to come through. The more data I can get, the better. Though it doesn't look like anyone wants to help, so this may all be a moot point anyway.

Do any of these variables impact dough as much as time or temp?  Individually, no, but, together, it all adds up.

You give me a 48 room temp NP dough, and, with a 'normal' formula, I can at the same temp, give you a dough that's ready in 24 or 96. I might be able to do it with just hydration (45% and 75%), but that could be a bit too ambitious.  Hydration + chlorinated water/non chlorinated + 0%/4% sugar + unmalted/malted flour + 1.75%/3% salt + fresh/1 year old yeast- well, that's a piece of cake.

Like I said, you'd have to go to some outlier formulations that nobody using this chart would be making with water you put bleach in and yeast you should have thrown away 3 years ago. Also like I said, I can spot an outlier and take it out of the data.

If you really believe you're right and you're not basing the whole thing on a red herring like making a cracker crust and a focaccia when neither is within the scope of the project, or using bleach water or old yeast, I propose a challenge:

I'll even give you most of the hydration range you want: 50-75%HR (if you can point me to a non-crackerish formula using 50%HR else the lowest regular pizza HR you can find), 2% +/- 2% sugar, 2% +/- 1% salt, tap or bottled water, malted or un-malted flour, and yeast (ADY or IDY) 1 year old or less – you pick and choose however you like. Let's see you come up with two doughs with a range of fermentation from 24 to 96 hours with the same amount of yeast and fermented at the exact same temperature. Pick the formulas, and I'll do the same experiment on this end, and we'll see who is right. One other thing - since your original objection was that the you couldn't trust submitted data because of the variations introduced by elements such as these, we have to assume working formulas. So whatever two formulas you come up with both must be able to make an American, NYish, or NPish pizza in the home oven - the dough must be able to be opened by hand, and it must eat at least OK when topped and baked.

What do you say?