Which is faster to type on – Torch or iPhone? And what does the difference cost?

Posted on August 20, 2010 by


This question has puzzled me for a while. I’ve always felt that my old Blackberry with a tactile QWERTY keyboard was much faster than anything I’ve used since, but receiving the new BlackBerry Torch 8900 prompted me to put the devices to the test.

If you want to know more about my impressions of the BlackBerry Torch itself, please see https://calvisblog.wordpress.com/2010/08/10/blackberry-torch-9800.

This is not an exhaustive scientific study – so I would strongly caution against using this data to substantiate or undermine any particular argument about the devices. However, I hope that what it will do is give you an indication of the potential impact of this particular issue: how fast can a user type on these mobile devices and what might that be costing.

As ever with these kind of results I think it is extremely important to be as fair as possible, so at the bottom I’ve included as many salient facts about the tests as possible so you can judge for yourself how relevant they are to your own situation.

In summary, the test consisted of a timing typing a short passage of text on to each device using the various input modes. The text I chose was from an actual email (81 words), as for me at least this is the most common kind of typing I do whilst mobile. I also tried each device in each input mode. So, let’s get to some result data:

Input mode
Orientation Technique Test
1 | 2 | 3
iPod Landscape Two thumbs 80| 69| 80 76
iPod Landscape Two hands/
one finger
65 | 75| 67 69
iPod Portrait Two thumbs 67 | 80 | 55 67
iPod Portrait Two hands/
one finger
55 | 53 | 56 55
iPod Portrait One hand 73 | 65 | 68 69
Torch Landscape Two thumbs 61 | 52 | 53 55
Torch Landscape Two hands/
one finger (2)
70 | 89 | 81 80
Torch Portrait Two thumbs (3) 91 | 83 | 97 90
Torch Portrait Two hands/
one finger
70 | 74 | 76 73
Torch Portrait One hand 115 | 95 | 98 103
Torch Physical Keyboard Two thumbs 57 | 48 | 50 52
Torch Physical Keyboard Two hands/
one finger
85 | 93 | 84 87
Torch Physical Keyboard One hand 90 | 92 | 97 93

Notes from table:

  1. Using two-thumbs when portrait I made more typos and hence the variability in timings to go back and correct.
  2. The smaller screen meant smaller buttons that made this test particularly error prone on the Torch. It is worth noting that BlackBerry appear to use optimizations in their landscape orientation that assume the two thumbs technique – you can see the impact of these optimizations in the results very clearly.
  3. This was even harder than two-thumbs on the iPod due to the even smaller buttons leading to many errors. The correction interface on the BlackBerry is not as fast as the iPod either, so these errors are costly.

Interpretation and commentary on tests

The chart below has the results summarized and ordered by fastest to slowest. I have included for comparison the time it took me to type the same text on my laptop keyboard.


There were some real surprises for me in these results. The fact that the Torch was noticeably faster in the two-thumb landscape mode I think is because its smaller screen actually makes this a little easier. Also, the time lost to corrections when trying to use two thumbs in portrait orientation on both devices surprised me.

I was also a little surprised to find how fast the two-hands/one-finger technique was (having imagined that a two thumb approach would be faster).

Unsurprising was the Torch physical keyboard coming out fastest, although note how closely it was followed by the iPod portrait two hand/one finger – just three seconds in it (which is nothing given the other variables in the test conditions).

It is interesting to note that the iPod results – coloured orange – are overall better than the Torch. My interpretation of that would be that the versatility of the iPod interface is significantly better than the Torch – but this is a point I’m sure you could argue any number of ways.

So, given the best scores on each devices, it is reasonable for me to conclude that there is no practical difference in typing speed between the Torch and the iPod/iPhone. Also, the screen keyboard on the Torch when used landscape with two-thumbs is just as fast. So, maybe my desire for the physical keyboard was ill founded and I should just give up my misconception that it is faster…?!

I also have a BlackBerry Storm, and its screen with “Sure Type” is not as fast as the screen only Torch, so I would expect (although have not tested) that it is not as fast as the Torch in landscape two-thumbs.

What does this cost?

Well, clearly between the fastest modes there is no difference. But what if there was? There are some many ways you could estimate the impact of these differences, but my method (determined before the test and so before I knew there wouldn’t be a difference!) followed this logic:

  • Dealing with email on the move has two benefits: (i) responsiveness to colleagues and (ii) saving time by dealing with email before returning to the offices;
  • Responsiveness undoubtedly has a value – in some cases a tangible value – but I can’t imagine a reliable and generic measure or estimate of this;
  • Saving time by dealing with email is the tangible benefit. To calculate this we need to look at the time taken to deal with an email in the office not on the mobile device, as it is the time saved when back at the office that we are measuring.

Based on this method, in my sample day I would have only dealt with half the email I actually did if I had used the least efficient typing technique, but the total saving in time of that for me when I get back to the office would be an irrelevant three minutes.

So the real advantage is to my colleagues – they get more of their emails dealt with while I am out of the office, and I don’t need to keep as much stuff in my head.

Does this mean it doesn’t really matter? No – not at all. Speed on the device is vital to me and allows me to keep on top of emerging opportunities or issues. I need an interface I am comfortable with and that keeps me productive. What I’ve learned though is that this could be either the iPhone or the Torch…

About the tests

I first analyzed a days’ use of my mobile. The day I chose was a day when I spent the whole day out of the office attending two fairly long client meetings, and a two and a half hour train journey at the start and end of the day. This meant I had three bursts of activity during the day – early morning, lunchtime, and afternoon/evening. There was nothing particularly unusual about this day that I am aware of in terms of my mobile email usage.

During this day, I sent twenty seven emails totalling 1,027 words and 5,550 characters (excluding spaces). That averages out at 38 words per email and 205 characters per email with an average word length of 5.4 characters. The text used for the test was a paragraph from an actual email I sent using my mobile. It contained 81 words and 423 characters – an average of 5.2 characters per word which is very close to my sample set. I do tend to write in full sentences on my mobile – particular of course for external e-mails – so most emails did not contain shorthand or abbreviations.

I use a BlackBerry for work use, but have only been using the Torch for a couple of weeks. I did the iPhone comparison using an iPod that I use daily (also for email) running iOS4. It would be fair to day that I do more typing of this kind of content on the BlackBerry than on the iPod, but I have certainly done a lot more typing on the iPod than on this particular BlackBerry.

It is also worth noting that I have small hands and fingers – undoubtedly an advantage on both of these devices!

The tests themselves consisted of me committing the text to memory and then typing it out once on each device using each input type. I repeated the test three times and took an average time. The test conditions were while on a moving train. This is a common place for me to use the device (so makes the results relevant to me at least!) but also was a compromise between doing the test standing still and walking – both of which I also do plenty of in reality.

I also tried using the various orientation/finger combinations, which are explained in the table below in more detail:

“Two thumbs” This is cradling the device in both hands and using both thumbs to do the typing (pretty much one thumb for each half of the keyboard)
“Two hands/one finger” This is holding the device in one hand and using the index finger or middle finger of the other hand to type.
“One hand/one thumb” This is using a single hand to both hold the device and type with the thumb of the same hand. I only find this practical in portrait modes on both devices, but is important to me because whilst on the move I usually have a briefcase in my other hand.

The input modes I tested are described below:

iPod I enabled keyboard auto-correction, auto-capitalization, caps lock and the “.” shortcut – and this is the configuration I am used to using on the iPod.The tests were conducted using the full keyboard in landscape and portrait orientations.
Torch I used the default settings on the Torch: word suggestion prediction, spelling suggestion, word learning context, word learning typing, names from contacts, *no* learn words from incoming email, and compound word suggestion order.I used the QWERTY layout and full keyboard in portrait orientation – again this is the configuration I am familiar with.
Posted in: Mobile Devices