Blind Faith Price Targets

[Johan]

This is the right sort of idea. All the curves converge to the longterm price target, LTPT = $3740 @ T = 2025-12-31. So given discount D, then the corresponding BFPT at time t in days is

BFPT = LTPT * (1 + D)^((T - t)/365).

So this is an exponential curve that passes through the LTPT.

Now the way we get a representive distribution of discounts is to use recent historical prices and solve for D. This is the idea of implied discount, the level of dicount to the LTPT that the market price implies. Where P(t) is the closing price on trading day t, the implied discount is

ID(t) =(LTPT/P(t))^(365/(T - t)) - 1.

So it is helpful compute and plot ID(t) for recent history. I routinely look at 2 or more years. This charts the swings in sentiment. As implied discount goes down it is a bull run, and when implied discount increases it is a bear run. It is easy to compute percentiles of the historical distribution. I use quartiles to set representative levels. One thing to contemplate as you look at look at a chart of historical implied discounts is how this distrubtion might change over time. The average discount presently is about 29%. Will this drift over time? In what direction? Yes it wiil change and will likely shrink over time. But the premise of the BFPT method is that this change in distribution will be quite slow, so that projecting out over just a few years is reasonable guage of variabikity in sentiment. Moreover, if the general trend is declining, this implies a bullish direction. Thus, the bias in this method is conservative for shareholders, that is, the method will likely understate future price distributions. I should point out, however, that this drift is very slow and the volatilty due to shifts in sentiment is so overwhelming as to make drift ignorable.

I hope this helps you explore this approach and am delighted that you are taking such an interest in it.

It makes a lot of sense when you paint that mathematical picture. What stands out to me is how though any and every information you can extract from such a curve is that it is 100% dependent on the price target (hence the Blind Faith Price Target). Change the price target and all the bear/bull/implied discount information get changed.

 

[jhm]

Thanks for this response. I also subscribe to the BFPT, and am interested in charting it. I am trying to work with the forumla you provided, but can't get it to do what I want. Could you provide the formula you use such that it could be plotted in excel? I am imagining rows with the date in it, and a formula that takes each date and inserts it into 't', with all other numbers in the calculation being provided as constants. Can you tell me what I'm doing wrong?

View attachment 86770
View attachment 86771

Everything looks correct. You will want to use a positive discount such as .285. You are using 0 discout which properly makes BFPT = LTPT. That is probably what is throwing you off.

BTW the LTPT I use is $3740. Of course, you can choose any value you want.

Good luck, jhm

 

[jhm]

Tesla BFPT Update

With all the crazy price movements lately, it seems to be a good time for a BFPT update to helps us keep our eyes on the long-term opportunity. It is curious that we were just pushing into the third quartile sentiment price of $287, but are push back to defending the median sentiment price at $266. Apparently the market is not ready to rush into fourth quartile bullish sentiment, but neither is it ready to really indulge in bearish sentiment either. This seems like a fair opportunity for long-term investors to buy at neutral sentiment prices.

On a technical note, I've expanded the training period of my BFPT from 12 months to 24 months. This dose not really change the median much, but expands the bearish tail a bit and gives the results a stronger historical footing with a little more conservatism around the downside potential.

Percentile	Implied Discount	7/22/2015	1/15/2016	7/22/2016	1/20/2017
0	32.82%	$193	$221	$256	$295
25	30.13%	$238	$271	$310	$354
50	28.76%	$266	$301	$343	$389
75	27.85%	$287	$323	$367	$415
100	25.47%	$349	$390	$438	$491

I'd be curious to hear how different people might frame an exit strategy around these BFPTs. At what price point, do you want to take some money off the table?

Best of luck, and as always make your own mistakes, not mine,

James

 

[austinEV]

With all the crazy price movements lately, it seems to be a good time for a BFPT update to helps us keep our eyes on the long-term opportunity. It is curious that we were just pushing into the third quartile sentiment price of $287, but are push back to defending the median sentiment price at $266. Apparently the market is not ready to rush into fourth quartile bullish sentiment, but neither is it ready to really indulge in bearish sentiment either. This seems like a fair opportunity for long-term investors to buy at neutral sentiment prices.

On a technical note, I've expanded the training period of my BFPT from 12 months to 24 months. This dose not really change the median much, but expands the bearish tail a bit and gives the results a stronger historical footing with a little more conservatism around the downside potential.

Percentile	Implied Discount	7/22/2015	1/15/2016	7/22/2016	1/20/2017
0	32.82%	$193	$221	$256	$295
25	30.13%	$238	$271	$310	$354
50	28.76%	$266	$301	$343	$389
75	27.85%	$287	$323	$367	$415
100	25.47%	$349	$390	$438	$491

I'd be curious to hear how different people might frame an exit strategy around these BFPTs. At what price point, do you want to take some money off the table?

Best of luck, and as always make your own mistakes, not mine,

James

I have been thinking about this quite a bit. That is, once you have a long term model (I have my own but I don't think it differs from BFPT materially) how does that translate into an investment? I had made myself this "holding table" which I have never followed:

I don't like it and don't think it is right but the table illustrates what we need. I think the concept is right. Based on how the stock is trading vs your favorite model, force yourself to adjust your portfolio so you don't get caught holding calls when it is overbought and you remember to take some leverage when it is oversold. I was going to take a few different models and actually back test a few simple recommendations to see what works the best. Ideally you would do a monte carlo analysis and see how sensitive it is to random fluctuations. The ideal table is probably pretty moderate.

 

[jhm]

I have been thinking about this quite a bit. That is, once you have a long term model (I have my own but I don't think it differs from BFPT materially) how does that translate into an investment? I had made myself this "holding table" which I have never followed:

View attachment 88224

I don't like it and don't think it is right but the table illustrates what we need. I think the concept is right. Based on how the stock is trading vs your favorite model, force yourself to adjust your portfolio so you don't get caught holding calls when it is overbought and you remember to take some leverage when it is oversold. I was going to take a few different models and actually back test a few simple recommendations to see what works the best. Ideally you would do a monte carlo analysis and see how sensitive it is to random fluctuations. The ideal table is probably pretty moderate.

I see the logic of your chart. You're basically increasing leverage to the degree that the stock is underpriced. What you may be missing--and this could be why you are not following this plan--is momentum. What do you really want to do in the second quartile? Well that depends on if the stock has been falling from the third quartile to second or rising from the first to second. If falling in to second, you might prefer buying stocks to buying calls because this minimizes your downside risk should momentum proceeds down to the first quartile. Conversely when ascending from first to second, this may be an excellent time to prefer buying calls to buying more stock. Basically you have both momentum and reversion to mean sentiment working in your favor, so leverage is good.

So I had been thinking I would sell my J16 calls when crossing from third to fourth quartiles. I failed to do this two days ago, and my calls have lost about a third of their value. They're worth holding onto at current price, but had I sold them Monday, I would be buying shares today and locking in longterm holdings. Oh well. I hesitated because I thought we would reach a new ATH soon. I still hold that view, but I need to be more careful about using too much leverage going above median sentiment to a new ATH.

 

[jhm]

New Long-Term Price Target

It's time for a BFPT update, and with the recent stock offering it is time to update the LTPT as well. Specifically the old LTPT was based on 125M shares, which just prior to the secondary should hAve been 127.1M, but with the secondary we are now at 130.2M. Assuming 4% annual dilution for the next ten years put longterm shares at 192.73M. I am making no change to the 2025 market cap of $691,980M. Thus, the new LTPT stands at $3590.45, down from $3739.82. Does this dilution make much difference in near term BFPTs? Not really, the 12 month median sentiment BFPT is now $348.67, down from $350.94, a difference of $2.27. So if that makes a difference in your investment strategy, you should not mess with Tesla.

So the good news is that if you want to accumulate Tesla shares whenever we are in the most bearish quartile, this is your buying opportunity. The first quartile price level for today is $247, and in 12 months at this level of sentiment the BFPT is $320. Moreover, as noted, the median sentiment BFPT is $349. So if you're willing to hold a year or more, there really is a lot of upside potential. I try not to let all this angst about not hitting a new ATH in the next couple of weeks cloud my longterm outlet. In 12 months, Tesla should see $349, and if you're holding stock it does not matter when in the next 12 months we first get there.

So here are the new price targets. Good luck, everyone.

Sentiment Percentile	Implied Discount	8/21/2015	1/15/2016	8/21/2016	1/20/2017	1/15/2018
0	32.37%	$196	$219	$260	$292	$385
5	30.97%	$219	$244	$287	$321	$419
25	29.45%	$247	$274	$320	$356	$459
50	28.27%	$272	$300	$349	$387	$494
75	27.39%	$292	$322	$372	$411	$522
95	25.87%	$330	$362	$416	$456	$575
100	25.01%	$355	$388	$444	$487	$607

 

[Gtoffo]

Hey JHM,

I'm trying to replicate your results but I'm having a hard time.

Using the 0 sentiment percentile the formula 3590×(1−0,32)^7 doesn't give me 219$.... what am I missing?

Also, how do you calculate the sentiment percentiles? We should create a shared spreadsheet.....

Right now I have this graph but the BFPTs are wrong....

 

[jhm]

Try this:

3,590.45÷(1+0.3237)^(9.967)
=219.4179991095

The number of years from 1/15/2016 to 12/31/2025 is just a little less than 10. I count calendar days and divide by 365 to get years.

Good luck.

 

[jhm]

One little comment on the current price. $219 is current price level at 5%. The opportunity to buy at sentiment this low only comes around a dozen times in a year. I added the 5% level because I wanted people to appreciate the opportunity.

Reversion to mean sentiment is so fast that the median BFPT is realistic in 12 months. That target is $349. So buying today near $219 really does give us a fair shot at $349 in 12 months. That is 59% upside. With that kind of upside potential, I find the leverage of options unneeded. The beauty of buying shares is that you can hold them forever, but with options you run the risk of the expiration date being ill timed and forcing you to take a loss. So with 59% upside within a year, this is an incredible opportunity to own shares for a lifetime.

I'm sure I must sound like a pitchman, but the whole BFPT methodology was built just to identify these opportunities as they arise. So it's here, and I'm excited to share the moment.

 

[MikeC]

One little comment on the current price. $219 is current price level at 5%. The opportunity to buy at sentiment this low only comes around a dozen times in a year. I added the 5% level because I wanted people to appreciate the opportunity.

Reversion to mean sentiment is so fast that the median BFPT is realistic in 12 months. That target is $349. So buying today near $219 really does give us a fair shot at $349 in 12 months. That is 59% upside. With that kind of upside potential, I find the leverage of options unneeded. The beauty of buying shares is that you can hold them forever, but with options you run the risk of the expiration date being ill timed and forcing you to take a loss. So with 59% upside within a year, this is an incredible opportunity to own shares for a lifetime.

I'm sure I must sound like a pitchman, but the whole BFPT methodology was built just to identify these opportunities as they arise. So it's here, and I'm excited to share the moment.

I'm with you. Great opportunity for long-term holders. Model X is around the freaking corner and the share price is where it was 18 months ago.

 

[jhm]

I'm with you. Great opportunity for long-term holders. Model X is around the freaking corner and the share price is where it was 18 months ago.

I do have the feeling that Tesla really grows in lumps. The Model S was a huge catalyst for growth and much of where the stock has gone over the last three years has just been the market digesting the potential of the Model S. So it expoded 3 years ago and has not really gone anywhere in the last 18 months. Tesla Energy is cool, but it has not yet really moved the stock, because sales of Powerpacks and Powerwalls have not yet moved the financials. So it is really important to own shares now before the Model X comes out because this will be the biggest driver of the stock for several years. I am hopeful that Tesla Energy will be a big hit by 2017. But the Model X really could be the biggest thing until Model 3, and most of those gains will be realized within the next 6 to 12 months.

So one of my own criticisms I'd level at my BFPT methology is that it assumes a very smooth value creation trajectory. It does not capture the lumpiness of how the business actually progresses. Someone building a detailed cash flow model can give expression to some non-linear growth dynamics, and so may be able to do a better job of navigating these near term events. But in defense of my super simple BFPT methology, it is transparent that it does not factor in these near term events. This gives us the ability to add our own commentary and mental adjustments to the simple trend. If we were actually above median sentiment right now, we'd have a hard time getting a solid buy signal out of this methodology. If we were at say $280, we'd be wresting with the idea that the median BFPT only shows a 25% upside to $349 in 12 month, but with the Model X coming out the upside is likely much greater than just 25%. Fortunately, we are spared that quandary by the fact that the stock is presently in quite bearish territory. The upside is big enough that we want to buy even if the Model X is not the second stage rocket engine. Frankly I have a hard time accounting for how we got so deep into bearish sentiment so quickly with all that is imminent. It may just be macro headwinds coupled with prelaunch jitters. Regardless, this is where the market puts us right now, and we can make what we like of this opportunity.

 

[Hogfighter]

I'm with you. Great opportunity for long-term holders. Model X is around the freaking corner and the share price is where it was 18 months ago.

The crux of this theory is that as each day ticks by, Tesla Motors is increasing in scale, scope, diversifying the product mix, and widening the competitive moat. Is Tesla Motors (the company) worth more today than last week? Of course...they are closer to Model X release, Tesla Energy product releases, GF expansion, to name only a few advances.

I for one have been happy to have had to opportunity to add to my portfolio with these prices. History will show whether we were correct (and rich) or incorrect (and broke).

 

[Gtoffo]

Well the formula I was using was wrong. But I should have realised that the multiplication didn't make any sense.

BTW I'm using the YEARFRAC formula to calculate precise year fractions (those 6th and 7th decimal points are CLEARLY crucial!)

Here is the new graph. I'm just missing the calculation about sentiment percentiles. I'm just using standard values that fit the top and bottom of the graph and then dividing up the interval. I can't come up with an easy way to do it properly. Do you have the algorithm posted somewhere?

I'm using quarterly intervals. So the latest drop isn't charted and the last 2 values are "random"

 

[jhm]

Well the formula I was using was wrong. But I should have realised that the multiplication didn't make any sense.

BTW I'm using the YEARFRAC formula to calculate precise year fractions (those 6th and 7th decimal points are CLEARLY crucial!)

Here is the new graph. I'm just missing the calculation about sentiment percentiles. I'm just using standard values that fit the top and bottom of the graph and then dividing up the interval. I can't come up with an easy way to do it properly. Do you have the algorithm posted somewhere?

I'm using quarterly intervals. So the latest drop isn't charted and the last 2 values are "random"

View attachment 92023

Is the red line the median curve or 75th percentile? With quarterly prices we don't see all the daily volatility which the curves are trying to account for. You might try putting daily close prices on this or at least quarterly highs and lows.

I'd also point out that the curves are calibrated in the recent 24 months. If you plot prices be for that when the stock hovered around $30, those prices will look very low relative to the curves. My interpretation is that no one had the idea at that time that Tesla could become a $700B market cap company in 10 to 15 years. That sort of ambition was not considered until the Model S came out and changed everything.

Good luck.

Edit. Oh, I looked more closely at your chart and saw the legend. You're basing this on discounts. So the red one is the median, which looks about right. In the most recent quarter it should near $270, so this looks a little high. You might check that this curve passes through the latest current levels I posted.

 

[Cattledog]

One little comment on the current price. $219 is current price level at 5%. The opportunity to buy at sentiment this low only comes around a dozen times in a year. I added the 5% level because I wanted people to appreciate the opportunity.

Reversion to mean sentiment is so fast that the median BFPT is realistic in 12 months. That target is $349. So buying today near $219 really does give us a fair shot at $349 in 12 months. That is 59% upside. With that kind of upside potential, I find the leverage of options unneeded. The beauty of buying shares is that you can hold them forever, but with options you run the risk of the expiration date being ill timed and forcing you to take a loss. So with 59% upside within a year, this is an incredible opportunity to own shares for a lifetime.

I'm sure I must sound like a pitchman, but the whole BFPT methodology was built just to identify these opportunities as they arise. So it's here, and I'm excited to share the moment.

Congratulations, nailed it.

 

[Gtoffo]

Is the red line the median curve or 75th percentile? With quarterly prices we don't see all the daily volatility which the curves are trying to account for. You might try putting daily close prices on this or at least quarterly highs and lows.

I'd also point out that the curves are calibrated in the recent 24 months. If you plot prices be for that when the stock hovered around $30, those prices will look very low relative to the curves. My interpretation is that no one had the idea at that time that Tesla could become a $700B market cap company in 10 to 15 years. That sort of ambition was not considered until the Model S came out and changed everything.

Good luck.

Edit. Oh, I looked more closely at your chart and saw the legend. You're basing this on discounts. So the red one is the median, which looks about right. In the most recent quarter it should near $270, so this looks a little high. You might check that this curve passes through the latest current levels I posted.

The idea was to have a long term overview of the methodology to understand how the stock moves to higher sentiment levels when some milestones are reached.
Big achievements reduce risk and consolidate the vision, therefore the discount rate should decrease in those cases.

Congratulations, nailed it.

I would caution short term traders not to put too much faith in this methodology however. Remember that those movements go both ways and there is no guarantee that when the stock is at a low sentiment level it isn't simply adjusting to a short/medium/long term downshift in sentiment. A week is an insufficient timeframe to determine if the buy signal from the method was correct.

This methodology should only be used by long terms stockholders I think.

With regards to the curves, I'm using the following formula: BFPT=LTPT/(1+D)^[(T-t)/365] and using the YEARFRAC formula to calculate precise year fractions. The red line correspond to the 0,2875 discount rate which is different from your median of 0,2827 (i've just divided in 4 the interval between 0,4 and 0,25....as I said, I'm not calculating percentiles as I can't come up with an easy way to do it, insight into your technique would be helpful).

Your analysis is based on the last 24 months..but what do you think is the appropriate discount rate for Tesla's LTPT today?


- - - Updated - - -

Switched to monthly data and added logarithmic version

 

[jhm]

Gtoffo,
The charts look great. The log scale is particularly helpful to see how these curves are constructed. They are actually straight lines in the log scale that converge at the LTPT. Thus, there is a contraction of volatility over time, and this is why it is important that the LTPT be well into the future, say 10 years or more.

I would second your caution about using this for short term trading. The point of buying when sentiment is really low is that you can lock in a really good price for a longterm investment. It is not any sort of signal for a near term reversal in prices. Short term price changes have almost nothing to do with the longterm value Tesla is building. So I use the BFPT methodology to see when current prices have lost meaningful connection with a longterm view.

I should also point out that I have done a regression of annual price changes on implied discounts. I did this as way to backtest the idea of buying at low sentiment (high implied discount) for a higher one-year return. So far what I have found is strong R-squared and that the expected return is roughly the return of the 12-month median sentiment BFPT to current price. Mean reversion seems to be fast enough for a 12 month time horizon. I would place a caveat here that we really do not have enough historical data for solid confirmation. But at least for now, I fail to detect anything that would invalidate this approach. It's always a good thing to test these things to make sure you are not doing something really stupid, even if you do not have enough data to prove you are doing something truly smart. That said, I am personally comfortable the 12-month median sentiment BFPT, so long as Tesla does not fundamentally change its growth trajectory.

If you want to compute quantiles of historical implied discounts, here is what you need. Get daily historical closing prices, say 24 months or so. For each historical closing price P_t on day t, compute implied discount,

ID_t = (LTPT / P_t)^(365/(T - t)) - 1.

Finally you can compute quantiles on the ID_t. In Excel, use the PERCENTILE function.

Additionally, for each daily historical price, I like to compute BFPT at that implied discount for various dates into the future. Then I can aggregate the targets any way I like, for example means and standard deviations. Quantiles of daily BFPTs are mathematically equivalent to BFPT of quantiles of implied discounts, which is one reason why I mostly work with quantiles.


Good luck implementing this for yourself. It really is the best way to really understand how this works and seeing potential for applications and improvements.

 

[Gtoffo]

Gtoffo,
The charts look great. The log scale is particularly helpful to see how these curves are constructed. They are actually straight lines in the log scale that converge at the LTPT. Thus, there is a contraction of volatility over time, and this is why it is important that the LTPT be well into the future, say 10 years or more.

I would second your caution about using this for short term trading. The point of buying when sentiment is really low is that you can lock in a really good price for a longterm investment. It is not any sort of signal for a near term reversal in prices. Short term price changes have almost nothing to do with the longterm value Tesla is building. So I use the BFPT methodology to see when current prices have lost meaningful connection with a longterm view.

I should also point out that I have done a regression of annual price changes on implied discounts. I did this as way to backtest the idea of buying at low sentiment (high implied discount) for a higher one-year return. So far what I have found is strong R-squared and that the expected return is roughly the return of the 12-month median sentiment BFPT to current price. Mean reversion seems to be fast enough for a 12 month time horizon. I would place a caveat here that we really do not have enough historical data for solid confirmation. But at least for now, I fail to detect anything that would invalidate this approach. It's always a good thing to test these things to make sure you are not doing something really stupid, even if you do not have enough data to prove you are doing something truly smart. That said, I am personally comfortable the 12-month median sentiment BFPT, so long as Tesla does not fundamentally change its growth trajectory.

If you want to compute quantiles of historical implied discounts, here is what you need. Get daily historical closing prices, say 24 months or so. For each historical closing price P_t on day t, compute implied discount,

ID_t = (LTPT / P_t)^(365/(T - t)) - 1.

Finally you can compute quantiles on the ID_t. In Excel, use the PERCENTILE function.

Additionally, for each daily historical price, I like to compute BFPT at that implied discount for various dates into the future. Then I can aggregate the targets any way I like, for example means and standard deviations. Quantiles of daily BFPTs are mathematically equivalent to BFPT of quantiles of implied discounts, which is one reason why I mostly work with quantiles.


Good luck implementing this for yourself. It really is the best way to really understand how this works and seeing potential for applications and improvements.

Thanks, I'll implement it next. Although graphically it's already pretty easy to see where the sentiment is compared to the last few months.

I wonder if Model X deliveries will push the discount down significantly. That would support the thesis that the stock is about to bounce significantly....

 

[jhm]

Thanks, I'll implement it next. Although graphically it's already pretty easy to see where the sentiment is compared to the last few months.

I wonder if Model X deliveries will push the discount down significantly. That would support the thesis that the stock is about to bounce significantly....

Yeah, it's a big question in my mind whether the Model X will actually shift the distribution. Certainly we should see some movement into more bullish price curves, but will that stick? Will median sentiment move to a lower implied discount or go back to about 28.5%? We did see that sort of shift from pre-Model S to post. You can see that quite clearly in your chart. So it could happen again. We shall see.

 

[MitchJi]

Reversion to mean sentiment is so fast that the median BFPT is realistic in 12 months. That target is $349. So buying today near $219 really does give us a fair shot at $349 in 12 months. That is 59% upside. With that kind of upside potential, I find the leverage of options unneeded.

Another point of view:
The BFPT timing model will apply equally well to timing LEAPS purchases. Two examples (missing the $219 opportunity) Friday Aug 28 close SP 248.48.

Strike	Last	Bid		Ask
180.00	90.00	87.70	91.50 
   68 of 90 (248 - 180) is intrinsic value 22 is time value, you pay a little less than 9% for 275% leverage
   
200.00	77.45	75.10	78.55
   48 of 78 (248 - 200) is intrinsic value 30 is time value, you pay a little less than 12% for 317% leverage

As you correctly infer, I roll my LEAPs well before expiration, 6-8 months depending on market conditions. I do this in order to minimize my theta losses (time decay of the option value) while leveraging my capital (when the stock price rises, the value of a call option goes up by more on a % basis).

If you roll those LEAPs 6-8 months before expiration, you will be paying less than half of the time values above (so under 6% for 317% leverage!), because the rate of time decay increases.

The beauty of buying shares is that you can hold them forever, but with options you run the risk of the expiration date being ill timed and forcing you to take a loss. So with 59% upside within a year, this is an incredible opportunity to own shares for a lifetime.

The worst case scenario is that TSLA is down for the entire 8 months before expiration (how likely is that?). That would cause a up to a 325% larger loss, but if you replaced the LEAPs with LEAPs at least you would pay a lower price for their replacements, and by the end of year two, assuming the increase is more than the first year loss you would be ahead.

NOTE: Trading in options definitely increases your risks. But the increased risk is pretty small compared to the potential upside. Using your estimate (a realistic 59% upside in 12 months) the potential upside is 310% (317% - 6%) x 59% profit is over 180%! This is not even taking into consideration that "when the stock price rises, the value of a call option goes up by more on a % basis".

Even if you roll the LEAPs into stock after one year, if the BFPT scenario works out you will come out way ahead!